The BASE SAS plus a bunch of functionalities I am missing in BASE SAS
- Package: BasePlus
- Version: 2.1.0
- Generated: 2024-09-09T12:55:04
- Author(s): Bartosz Jablonski ([email protected]), Quentin McMullen ([email protected])
- Maintainer(s): Bartosz Jablonski ([email protected])
- License: MIT
- File SHA256:
F*DFA83F8E0D7424DEB63D49620392068BC68D766552E2804CB6B01DE8E5A87769for this version - Content SHA256:
C*AB16F5B6538515607C3C254E401DC1ACA7293AB36990227F6E7792145CEEAB87for this version
The BasePlus package implements useful functions and functionalities I miss in the BASE SAS.
It is inspired by various people, e.g.
- at the SAS-L discussion list
- at the communities.sas.com (SASware Ballot Ideas)
- at StackOverflow
- at the Office...
- etc.
Kudos to all who inspired me to generate this package: Mark Keintz, Paul Dorfman, Richard DeVenezia, Christian Graffeuille, Allan Bowe, Anamaria Calai, Michal Ludwicki, Quentin McMullen, Kurt Bremser, Leonid Batkhan, Louise Hadden.
Example 1: One-dimensional array functions. Array parameters to subroutine calls must be 1-based.
data _null_;
array X[4] _temporary_ (. 1 . 2);
call arrMissToRight(X);
do i = 1 to 4;
put X[i]= @;
end;
put;
call arrFillMiss(17, X);
do i = 1 to 4;
put X[i]= @;
end;
put;
call arrFill(42, X);
do i = 1 to 4;
put X[i]= @;
end;
run;Example 2: Delete dataset by name.
data toDrop;
x = 17;
run;
data _null_;
p = delDataset("toDrop");
put p=;
run;Example 3: Strings concatenation with format.
data test;
x = 1 ; y = . ; z = 3 ;
t = "t"; u = " "; v = "v";
array a[*] x y z;
array b[*] t u v;
length s1 s2 s3 s4 $ 17;
s1 = catXFn("z5.", "#", A);
s2 = catXFi("z5.", "#", A);
s3 = catXFc("upcase.", "*", B);
s4 = catXFj("upcase.", "*", B);
put (_all_) (=);
run;Example 4: Useful formats.
data _null_;
input x @@;
put @1 x= @11 x= bool. @21 x= int. @31 x= ceil. @41 x= floor.;
cards;
. ._ .A -10 -3.14 0 3.14 10
;
run;Example 5: Getting variables names from datasets.
%put *%getVars(sashelp.class
,pattern = ght$
,sep = +
,varRange = _numeric_)*;Example 6: Quick sort as an alternative to call sortn()
data _null_;
array test[25000000] _temporary_ ;
t = time();
call streaminit(123);
do _N_ = 25000000 to 1 by -1;
test[_N_] = rand("uniform");
end;
t = time() - t;
put "Array population time: " t;
t = time();
call quickSortLight (test);
t = time()-t;
put "Sorting time: " / t=;
run;Example 7: De-duplicate values from a space separated list.
%let list = 4 5 6 1 2 3 1 2 3 4 5 6;
%put *%dedupListS(&list.)*;Example 8: Zip elements of two space separated list.
%let x = %zipEvalf(1 2 3 4 5 6, 2018 2019 2020, argMd=5, function=MDY, format=date11.);
%put &=x;Example 9: Simple Rain Cloud plot.
%rainCloudPlot(sashelp.cars,DriveTrain,Invoice)The output can be seen in the md file.
Example 10: Zip SAS library.
%zipLibrary(sashelp, libOut=work)
%unzipLibrary(%sysfunc(pathname(work)), zip=sashelp, mode=S, clean=1)Example 11: Long dataset names.
data %LDSN( work. peanut butter & jelly with a "Hot-Dog" in [a box], popcorn, and s*t*a*r*s (drop = sex rename=(name=first_name) where = (age in (12,13,14))) );
set sashelp.class;
run;
proc print data = %LDSN( work. peanut butter & jelly with a "Hot-Dog" in [a box], popcorn, and s*t*a*r*s );
run;
data MyNextDataset;
set %LDSN( work. peanut butter & jelly with a "Hot-Dog" in [a box], popcorn, and s*t*a*r*s );
where age > 12;
run;Example 12: List, to the log, content of home directory.
%bpPIPE(ls -la ~/)EXAMPLE 13 Get list of all files and directories from C:\SAS_WORK\:
%dirsAndFiles(C:\SAS_WORK\,ODS=work.result)EXAMPLE 14 Text repetition:
%put %repeatTxt(#,15,s=$) HELLO SAS! %repeatTxt(#,15,s=$);EXAMPLE 15 Integer list:
%put %intsList(42);EXAMPLE 16 Split dataset into blocks of 5 observations:
%splitDSIntoBlocks(5, sashelp.class, classBlock)EXAMPLE 17 Split dataset into 7 parts:
%splitDSIntoParts(7, sashelp.cars, carsPart)EXAMPLE 18 Return path to temporary file:
filename f temp;
%put %filePath(f);EXAMPLE 19 Get titles:
title1 j=c "Hi Roger" ;
title2 j=l "Good Morning" ;
title3 "How are you?" ;
title4 ;
title5 "Bye bye!" ;
%put %GetTitle(1 2 3 5, dlm=s, qt='') ;EXAMPLE 20 Format and informat macro variables values:
%put %fmt(12345, date9.) %fmt(12345, yymmdd10.);
%put %infmt($111234, dollar10.2);
%put %infmt($111.234, dollar10.2);
%let text = ##%fmt(ABC, $char9., -C)##;
%put &text.;EXAMPLE 21 "Macro including" a text file:
filename f "%workpath()/testFile1.txt";
data _null_;
file f;
put "13 14 15";
run;
data testDataset;
set sashelp.class;
where age in ( %mInclude(f) );
run;EXAMPLE 22 Repeating texts and lists:
options mprint;
data work.A;
x=17;
data work.B;
x=42;
data work.C;
x=303;
run;
data work.test5;
set
%repeatTxt(work.A work.B work.C, 5)
;
run;
data Times2_A3B4C5;
set
%repList(work.A work.B work.C, times = 2, each = 3 4 5)
;
run;EXAMPLE 23 Date and time one-liners:
%put #%today()#%date()#%time()#%datetime()#;
%put @%today(yymmdd10.)@%date(date11.)@%time(time8.)@%datetime(e8601dt.)@; EXAMPLE 24 Months shifting:
%put
Past: %monthShift(2023, 1, -1)
Current: %monthShift(2023, 1 )
Future: %monthShift(2023, 1, +1)
;EXAMPLE 25 Zipping and unzipping directories:
options dlCreateDir;
libname arch1 "%workPath()/testArch1";
libname arch2 "%workPath()/testArch2";
filename arch1 "%workPath()/testArch1";
data _null_;
file arch1(test1.txt);
put "text for test file 1";
data _null_;
file arch1(test2.txt);
put "text for test file 2";
data _null_;
file arch1(test3.txt);
put "text for test file 3";
run;
data arch1.class(index=(name));
set sashelp.class;
run;
data arch1.cars(index=(model));
set sashelp.cars;
run;
%zipArch(
archName2.zip
, pathRef = arch1
, target = %workPath()/testArch2
, list = 1
, overwrite = 1
)
%unzipArch(
archName2.zip
, path = %workPath()/testArch2
, target = %workPath()/testArch2
, clean=1
, list=1
);
EXAMPLE 26 Downloading data from the internet to a local directory:
%downloadFilesTo(~/directoryA)
datalines4;
https://www.lexjansen.com/wuss/2023/WUSS-2023-Paper-189.pdf
https://www.lexjansen.com/wuss/2023/WUSS-2023-Paper-189.zip
;;;;
run;EXAMPLE 27 Conditional value assignment:
%let x = A B C;
%let y = %iffunc((%scan(&x.,1)=A),Starts with "A"., Does not start with "A".);
%put &=y.;EXAMPLE 28 Converting variables names to lowercases:
data a1 a2 a3 b_x b_y b_z;
set sashelp.class(obs=1);
run;
%put #%expandDataSetsList(lib=work,datasets=a1-a3 b_:)#;
proc print data=a1;
proc print data=b_x;
run;
%unifyVarsCaseSize(work,a1-a3 b_:)
proc print data=a1;
proc print data=b_x;
run;Package contains additional content, run: %loadPackageAddCnt(BasePlus) to load it
or look for the baseplus_AdditionalContent directory in the packages fileref
localization (only if additional content was deployed during the installation process).
SAS package generated by SAS Package Framework, version 20240711
The BasePlus package consists of the following content:
The bpPIPE() [Base Plus PIPE] macro executes OS command and print to the log output of the execution.
Under the hood it uses _ filename reference to PIPE device.
The basic syntax is the following, the <...> means optional parameters:
%bpPIPE( <OS command goes here> )Arguments description:
- NO Arguments - Everything inside brackets is treated as an OS command.
EXAMPLE 1. List, to the log, content of D and C drives:
%bpPIPE(D: & dir & dir "C:\")EXAMPLE 2. List, to the log, content of home directory:
%bpPIPE(ls -halt ~/)The %dedupListC() macro deletes duplicated values from
a COMMA-separated list of values. List, including separators,
can be no longer than a value carried by a single macro variable.
Returned value is unquoted. Leading and trailing spaces are ignored.
The %dedupListC() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%dedupListC(
list,of,comma,separated,values
)Arguments description:
list- A list of comma-separated values.
EXAMPLE 1. Basic use-case one. Delete duplicated values from a list.
%put *%dedupListC(a,b,c,b,c)*;
%put *%dedupListC(a,b c,b c)*;
%put *%dedupListC(%str(a,b,c,b,c))*;
%put *%dedupListC(%str(a),%str(b),%str(c),b,c)*;EXAMPLE 2. Leading and trailing spaces are ignored. Delete duplicated values from a list.
%put *%dedupListC( a , b b , c , b b, c )*;EXAMPLE 3. Macro variable as an argument. Delete duplicated values from a list.
%let list = 4, 5, 6, 1, 2, 3, 1, 2, 3, 4, 5, 6;
%put *%dedupListC(&list.)*;The %dedupListP() macro deletes duplicated values from
a PIPE(|)-separated list of values. List, including separators,
can be no longer than a value carried by a single macro variable.
Returned value is unquoted. Leading and trailing spaces are ignored.
The %dedupListP() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%dedupListP(
list|of|pipe|separated|values
)Arguments description:
list- A list of pipe-separated values.
EXAMPLE 1. Basic use-case one. Delete duplicated values from a list.
%put *%dedupListP(a|b|c|b|c)*;
%put *%dedupListP(a|b c|b c)*;
%put *%dedupListP(%str(a|b|c|b|c))*;
%put *%dedupListP(%str(a)|%str(b)|%str(c)|b|c)*;EXAMPLE 2. Leading and trailing spaces are ignored. Delete duplicated values from a list.
%put *%dedupListP( a | b b | c | b b| c )*;EXAMPLE 3. Macro variable as an argument. Delete duplicated values from a list.
%let list = 4|5|6|1|2|3|1|2|3|4|5|6;
%put *%dedupListP(&list.)*;The %dedupListS() macro deletes duplicated values from
a SPACE-separated list of values. List, including separators,
can be no longer than a value carried by a single macro variable.
Returned value is unquoted.
The %dedupListS() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%dedupListS(
list of space-separated values
)Arguments description:
list- A list of space-separated values.
EXAMPLE 1. Basic use-case one. Delete duplicated values from a list.
%put *%dedupListS(a b c b c)*;
%put *%dedupListS(a b,c b,c)*;
%put *%dedupListS(%str(a b c b c))*;
%put *%dedupListS(%str(a) %str(b) %str(c) b c)*;EXAMPLE 2. Macro variable as an argument. Delete duplicated values from a list.
%let list = 4 5 6 1 2 3 1 2 3 4 5 6;
%put *%dedupListS(&list.)*;The %dedupListX() macro deletes duplicated values from
a X-separated list of values, where the X represents
a single character separator. List, including separators,
can be no longer than a value carried by a single macro variable.
Caution. The value of X has to be in the first byte of the list,
just after the opening bracket, i.e. (X...).
Returned value is unquoted. Leading and trailing spaces are ignored.
The %dedupListX() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%dedupListX(
XlistXofXxXseparatedXvalues
)Arguments description:
list- A list of X-separated values.
EXAMPLE 1. Basic use-case one. Delete duplicated values from a list.
%put *%dedupListX(|a|b|c|b|c)*;
%put *%dedupListX( a b c b c)*;
%put *%dedupListX(,a,b,c,b,c)*;
%put *%dedupListX(XaXbXcXbXc)*;
%put *%dedupListX(/a/b/c/b/c)*;
data _null_;
x = "%dedupListX(%str(;a;b;c;b;c))";
put x=;
run;EXAMPLE 2. Leading and trailing spaces are ignored. Delete duplicated values from a list.
%put *%dedupListX(| a | b.b | c | b.b| c )*;
%put *%dedupListX(. a . b b . c . b b. c )*;EXAMPLE 3. Macro variable as an argument. Delete duplicated values from a list.
%let list = 4$5.5$6$1$2$3$1$2$3$4$5.5$6;
%put *%dedupListX($&list.)*;
%let list = 4$ 5.5$ 6$ 1$ 2$ 3$ 1$ 2$ 3$ 4$ 5.5$ 6$;
%put *%dedupListX( &list.)*;The %dirsAndFiles() macro allows to extract info about all files
and subdirectories of a given root directory.
The extracted info may be just a list of files and subdirectories or, if
the details= parameter is set to 1, additional operating system information
is extracted (information is OS-dependent and gives different results for Linux
and for Windows)
The extracted info can be narrowed down to files (keepFiles=1) or to
directories (keepDirs=1) if need be.
The extracted info can be presented in wide or long format (longFormat=1).
The extracted info for files can be narrowed down to only files with particular
extension, for example: fileExt=sas7bdat.
The extracted info can be narrowed down maximal path depth
by setting up the maxDepth= parameter.
See examples below for the details.
The macro is based on Kurt Bremser's "Talking to Your Host" article presented at WUSS 2022 conference.
The article is available here and also as an additional content of this package. The paper was awarded the "Best Paper Award - Programming".
The basic syntax is the following, the <...> means optional parameters:
%dirsAndFiles(
root
<,ODS=>
<,details=>
<,keepDirs=>
<,keepFiles=>
<,longFormat=>
<,fileExt=>
<,maxDepth=>
)Arguments description:
root- Required, path to be searched for information.
-
ODS=work.dirsAndFilesInfo- Optional, output data set, name of a dataset to store information. -
details=0- Optional, indicates if detailed info will be collected,1= yes,0= no. -
keepDirs=1- Optional, indicates if directories info will be collected,1= yes,0= no. -
keepFiles=1- Optional, indicates if files info will be collected,1= yes,0= no. -
longFormat=0- Optional, indicates if output be in long format,1= yes,0= no. -
fileExt=- Optional, if not missing then indicates a list of space-separated file extensions to filter out results. -
maxDepth=0- Optional, if not zero then indicates maximum depth of search in the root path.
EXAMPLE 1. Get list of files and directories:
%dirsAndFiles(C:\SAS_WORK\,ODS=work.result1)EXAMPLE 2. Get detailed info:
%dirsAndFiles(C:\SAS_WORK\,ODS=work.result2,details=1)EXAMPLE 3. Get only files info:
%dirsAndFiles(C:\SAS_WORK\,ODS=work.result3,keepDirs=0)
%dirsAndFiles(C:\SAS_WORK\,ODS=work.result5,keepDirs=0,details=1)EXAMPLE 4. Get only directories info:
%dirsAndFiles(C:\SAS_WORK\,ODS=work.result4,keepFiles=0)
%dirsAndFiles(C:\SAS_WORK\,ODS=work.result6,keepFiles=0,details=1)EXAMPLE 5. Filter out by sas extension:
%dirsAndFiles(~/,ODS=work.result7,fileExt=sas)
%dirsAndFiles(~/,ODS=work.result8,fileExt=sas,details=1)EXAMPLE 6. Keep result in the long format:
%dirsAndFiles(~/,ODS=work.result9,details=1,longFormat=1)EXAMPLE 7. Get info for maximum depth of 2:
%dirsAndFiles(C:\SAS_WORK\,ODS=work.result10,details=1,maxDepth=2)EXAMPLE 8. How locked/unavailable files are handled:
%dirsAndFiles(%sysfunc(pathname(WORK)),ODS=work.result11,details=1)EXAMPLE 9. Not existing directory:
%dirsAndFiles(%sysfunc(pathname(WORK))/noSuchDir,ODS=work.result12,details=1)The functionExists() macro function tests
if given funcion exists in the SAS session.
The sashelp.vfunc view is used.
See examples below for the details.
The %functionExists() macro executes like a pure macro code.
The function is a result of cooperation with Allan Bowe
The basic syntax is the following, the <...> means optional parameters:
%functionExists(
funName
)Arguments description:
funName- Required, the name of the function existence of which you are testing.
EXAMPLE 1. Test if function exists:
%put %functionExists(HASHING);
%put %functionExists(COSsinLOG);The getVars() and QgetVars() macro functions allow to extract variables names form a dataset according to a given pattern into a list.
The getVars() returns unquoted value [by %unquote()]. The QgetVars() returns quoted value [by %superq()].
See examples below for the details.
The %getVars() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%getVars(
ds
<,sep=>
<,pattern=>
<,varRange=>
<,quote=>
<,mcArray=>
<,ignoreCases>
)Arguments description:
ds- Required, the name of the dataset from which variables are to be taken.
-
sep = %str( )- Optional, default value%str( ), a variables separator on the created list. -
pattern = .*- Optional, default value.*(i.e. any text), a variable name regexp pattern, by default case INSENSITIVE! -
varRange = _all_- Optional, default value_all_, a named range list of variables. -
quote =- Optional, default value is blank, a quotation symbol to be used around values. -
mcArray=- Optional, default value is blank. 1) When null - the macro behaves like a macro function and returns a text string with variables list. 2) When not null - behaviour of the macro is altered. In such case a macro array of selected variables, named withmcArrayvalue as a prefix, is created. Furthermore a macro named asmcArrayvalue is generated. (see the macroArray package for the details). WhenmcArray=parameter is active thegetVarsmacro cannot be called within the%putstatement. Execution like:%put %getVars(..., mcArray=XXX);will result with
an Explicit & Radical Refuse Of Run (aka ERROR). -
ignoreCases=- Optional, default value is 1.
Indicates if search should be case insensitive.
EXAMPLE 1. A list of all variables from the sashelp.class dataset:
%put *%getVars(sashelp.class)*;EXAMPLE 2. A list of all variables from the sashelp.class dataset separated by backslash:
%let x = %getVars(sashelp.class, sep=\);
%put &=x;EXAMPLE 3. Use of regular expressions: a) A list of variables which name contains "i" or "a"
%put *%getVars(sashelp.class, pattern=i|a)*;b) A list of variables which name starts with "w"
%put *%getVars(sashelp.class, pattern=^w)*;c) A list of variables which name ends with "ght"
%put *%getVars(sashelp.class, pattern=ght$)*;EXAMPLE 4. A list of numeric variables which name starts with "w" or "h" or ends with "x"
%put *%getVars(sashelp.class, sep=+, pattern=^(w|h)|x$, varRange=_numeric_)*;EXAMPLE 5.
data test;
array x[30];
array y[30] $ ;
array z[30];
run;a) A list of variables separated by a comma:
%put *%getVars(test, sep=%str(,))*;b) A list of variables separated by a comma with suffix 5 or 7:
%put *%getVars(test, sep=%str(,), pattern=(5|7)$)*;c) A list of variables separated by a comma with suffix 5 or 7 from a given variables range:
%put *%getVars(test, sep=%str(,), varRange=x10-numeric-z22 y6-y26, pattern=(5|7)$)*;EXAMPLE 6. Case of quotes and special characters when the quote= parameter is not used:
a) one single or double qiote:
%put *%bquote(%getVars(sashelp.class, sep=%str(%")))*;
%put *%bquote(%getVars(sashelp.class, sep=%str(%')))*;b) two single or double qiotes:
%put *"%bquote(%getVars(sashelp.class,sep=""))"*;
%put *%str(%')%bquote(%getVars(sashelp.class,sep=''))%str(%')*;c) comma-separated double quote list:
%put *"%getVars(sashelp.class,sep=%str(", "))"*;d) comma-separated single quote list:
%put *%str(%')%getVars(sashelp.class,sep=', ')%str(%')*;
%let x = %str(%')%getVars(sashelp.class,sep=', ')%str(%');
%put *%str(%')%QgetVars(sashelp.class,sep=', ')%str(%')*;
%let y = %str(%')%QgetVars(sashelp.class,sep=', ')%str(%');
%let z = %unquote(&y.);e) ampersand (&) as a separator [compare behaviour]:
%put *%getVars(sashelp.class,sep=&)*;
%let x = %getVars(sashelp.class,sep=&);
%put *%getVars(sashelp.class,sep=%str( & ))*;
%let x = %getVars(sashelp.class,sep=%str( & ));
%put *%QgetVars(sashelp.class,sep=&)*;
%let y = %QgetVars(sashelp.class,sep=&);
%let z = %unquote(&y.);
%put *%QgetVars(sashelp.class,sep=%str( & ))*;
%let y = %QgetVars(sashelp.class,sep=%str( & ));
%let z = %unquote(&y.);
%put *%getVars(sashelp.class,sep=&)*;
%let x = %getVars(sashelp.class,sep=&);
%put *%getVars(sashelp.class,sep=%str( & ))*;
%let x = %getVars(sashelp.class,sep=%str( & ));f) percent (%) as a separator [compare behaviour]:
%put *%QgetVars(sashelp.class,sep=%)*;
%let y = %QgetVars(sashelp.class,sep=%);
%let z = %unquote(&y.);
%put *%QgetVars(sashelp.class,sep=%str( % ))*;
%let y = %QgetVars(sashelp.class,sep=%str( % ));
%let z = %unquote(&y.);EXAMPLE 7. Case of quotes and special characters when the quote= parameter is used:
a) one single or double qiote:
%put *%getVars(sashelp.class, quote=%str(%"))*;
%put *%getVars(sashelp.class, quote=%str(%'))*;b) two single or double quotes:
%* this gives an error: ;
%* %put *%getVars(sashelp.class,quote="")*;
%* %put *%getVars(sashelp.class,quote='')*;
%* this does not give an error: ;
%put *%QgetVars(sashelp.class,quote="")*;
%put *%QgetVars(sashelp.class,quote='')*;c) comma-separated double quote list:
%put *%getVars(sashelp.class,sep=%str(,),quote=%str(%"))*;d) comma-separated single quote list:
%let x = %getVars(sashelp.class,sep=%str(,),quote=%str(%'));
%put &=x.;EXAMPLE 8. Variables that start with A and do not end with GHT:
data class;
set sashelp.class;
Aeight = height;
run;
%put *%getVars(class, pattern = ^A(.*)(?<!ght)$, quote=%str(%"))*;EXAMPLE 9. Variables that do not start with N and do not end with GHT:
data class;
set sashelp.class;
Aeight = height;
Neight = height;
run;
%put *%getVars(class, pattern = ^(?!N.*)(.*)(?<!ght)$, quote=%str(%"))*;EXAMPLE 10. Composition with itself:
data class;
set sashelp.class;
Age_C = put(Age, best32.);
Height_C = put(Height, best32.);
Weight_C = put(Weight, best32.);
run;
%put #%getVars(class, varRange=_numeric_, sep=%str(: ))# <- no : at the end!!;
%put #%getVars(class, varRange=%getVars(class, varRange=_numeric_, sep=%str(: )):, sep=\)#;EXAMPLE 11. Create a macro array XYZ... of variables names and an additional
macro %XYZ() which allows easy access to the list. Can be used with
the %do_over() macro (provided with the macroArray package).
data test;
array x[30];
array y[30] $ ;
array z[30];
run;
%getVars(test
,mcArray=XYZ
,varRange=x10-numeric-z22 y6-y26
,pattern=(5|7)$
,quote=#)
%put _user_;
%put *%XYZ(1)**%XYZ(2)*%XYZ(3)*;
%* Load the macroArray package first. ;
%put %do_over(XYZ);The intsList() macro function allows to print a list of
integers starting from start up to end incremented by by
and separated by sep=.
If start, end or by are non-integers the are converted to integers.
See examples below for the details.
The %intsList() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%intsList(
start
<,end>
<,by>
<,sep=>
)Arguments description:
-
start- Required, the first value of the list. Ifendis missing then the list is generated from 1 tostartby 1. -
end- Required/Optional, the last value of the list. -
by- Required/Optional, the increment of the list. If missing then set to1. Cannot be equal to0.
s = %str( )- Optional, it is a separator between elements of the list. Default value is space.
EXAMPLE 1. Simple list of integers from 1 to 10 by 1:
%put %intsList(10);EXAMPLE 2. Ten copies of sashelp.class in test11 to test20:
data
%zipEvalf(test, %intsList(11,20))
;
set sashelp.class;
run;EXAMPLE 3. Non-integers are converted to integers, the list is 1 3 5:
%put %intsList(1.1,5.2,2.3);EXAMPLE 4. A list with a separator:
%put %intsList(1,5,2,sep=+);The LDSN (Long DataSet Names) macro function allows to use an "arbitrary" text string to name a dataset.
The LDSN macro has some limitation described below, to overcome them another macro, with different name: LDSNM (Long DataSet Names Modified) was created. See its description to learn how to use it.
The idea for the macro came from the following story:
Good friend of mine, who didn't use SAS for quite some time, told me that he lost a few hours for debugging because he forgot that the SAS dataset name limitation is 32 bytes.
I replied that it shouldn't be a problem to do a workaround
for this inconvenience with a macro and the MD5() hashing function.
I said: The macro should take an "arbitrary string" for a dataset
name, convert it, with help of MD5(), to a hash digest, and
create a dataset with an "artificial" hex16. formated name.
Starting with something like this:
data %LDSN(work. peanut butter & jelly with a hot-dog in [a box] and s*t*a*r*s (drop = sex rename=(name=first_name) where = (age in (12,13,14))) );
set sashelp.class;
run;the macro would do:
%sysfunc(MD5(peanut butter & jelly with a hot-dog in [a box] and s*t*a*r*s), hex16.)and (under the hood) return and execute the following code:
data work.DSN_41D599EF51FBA58_(drop = sex rename=(name=first_name) where = (age in (12,13,14))) ;
set sashelp.class;
run;Also in the next data step user should be able to do:
data my_next_data_step;
set %DSN(work. peanut butter & jelly with a hot-dog in [a box] and s*t*a*r*s);
run;and work without the "dataset-name-length-limitation" issue.
See examples below for the details.
The %LDSN() macro executes like a pure macro code.
Known "Limitations":
-
dataset name cannot contain dots (
.) since they are used as separators! -
dataset name cannot contain round brackets(
(and)) since they are used as separators (but[]and{}are allowed)! -
dataset name cannot contain unpaired quotes (
'and"), text:a "hot-dog"is ok, butJohn's dogis not!
Behaviour:
-
dataset name text is converted to upcase
-
dataset name text leading and trailing spaces are ignored, e.g. the following will give the same hash digest:
%ldsn(work.test),%ldsn( work.test),%ldsn(work.test ),%ldsn(work .test),%ldsn(work. test),%ldsn(work . test). -
macro calls of the form:
data %LDSN(); run;,data %LDSN( ); run;,data %LDSN( . ); run;or evendata %LDSN( . (keep=x)); run;are resolved to empty string, so the result is equivalent todata; run;
The basic syntax is the following, the <...> means optional parameters:
%LDSN(
arbitrary text string (in line with limitations)
)The text string is concider as "fully qualified dataset name", i.e. macro
assumes it may contain library as prefix and data set options as sufix.
See the %LDsNm() macro for comparison.
EXAMPLE 1.
options nomprint source nomlogic nosymbolgen ls = max ps = max;
data %LDSN( work. peanut butter & jelly with a "Hot-Dog" in [a box], popcorn, and s*t*a*r*s (drop = sex rename=(name=first_name) where = (age in (12,13,14))) );
set sashelp.class;
run;
proc print data = %LDSN( work. peanut butter & jelly with a "Hot-Dog" in [a box], popcorn, and s*t*a*r*s );
run;
data MyNextDataset;
set %LDSN( work. peanut butter & jelly with a "Hot-Dog" in [a box], popcorn, and s*t*a*r*s );
where age > 12;
run;The LDSNM (Long DataSet Names Modified) macro function allows to use an "arbitrary" text string to name a dataset.
The LDSN macro had some limitation (see its documentation), to overcome them
another %LDSNM() (Long DataSet Names Modified) macro was created.
The main idea behind the %LDSNM() is the same as for %LDSN() - see the description there.
The %LDSNM() works differently then the %LDSN().
The %LDSN() assumed that both libname and dataset options could
be passed as elements in macro argument, e.g.
data %LDSN( WORK.peanut butter & jelly with a hot-dog in [a box] and s*t*a*r*s (drop = sex) );
set sashelp.class;
run;The %LDSNM(), in contrary, assumes that both libname and dataset options are
passed outside the macro, i.e.
data WORK.%LDSNM( peanut butter & jelly with a hot-dog in [a box] and s*t*a*r*s ) (drop = sex);
set sashelp.class;
run;This approach reduces some limitations the LDSN has.
The additional feature of the %LDSNM() is that when the macro is called
a global macro variable, which name is the same as hashed dataset name, is created.
The macro variable value is the text of the argument of the macro. For example
the following macro call:
data %LDSNM(John "x" 'y' dog);
set sashelp.class;
where name = 'John';
run;creates DSN_BF1F8C4D6495B34A_ macro variable with value: JOHN "X" 'Y' DOG.
The macro variable is useful when combined with symget() function and
the indsname= option to get the original text string value back,
like in this example:
data test;
set %LDSNM(John "x" 'y' dog) indsname = i;
indsname = symget(scan(i,-1,"."));
run;See examples below for the details.
The %LDSN() macro executes like a pure macro code.
Known "Limitations":
-
dataset name cannot contain unpaired round brackets(
(and)) (but unmatched[]and{}are allowed)! -
dataset name cannot contain unpaired quotes (
'and"), text:a "hot-dog"is ok, butJohn's dogis not!
Behaviour:
-
dataset name text is converted to upcase
-
dataset name text leading and trailing spaces are ignored, e.g. the following will give the same hash digest:
%ldsn(test),%ldsn( test),%ldsn(test ). -
macro calls of the form:
data %LDSN(); run;ordata %LDSN( ); run;are resolved to empty string, so the result is equivalent todata; run; -
created macro variable is global in scope.
The basic syntax is the following, the <...> means optional parameters:
%LDSNM(
arbitrary text string (in line with limitations)
)The text string is consider as "only dataset name", i.e. macro does not
assume it contain library as prefix or data set options as suffix.
See the %LDSN() macro for comparison.
EXAMPLE 1.
data %LDSNM(John "x" 'y' & dog);
set sashelp.class;
where name = 'John';
run;
data %LDSNM(John "x"[ 'y' & dog);
set sashelp.class;
where name = 'John';
run;
data %LDSNM(John "x" 'y'} & dog);
set sashelp.class;
where name = 'John';
run;EXAMPLE 2.
data work.%LDsNm( peanut butter & jelly, a hot-dog in [a box], and s(*)t(*)a(*)r(*)s!! ) (drop = sex rename=(name=first_name) where = (age in (12,13,14)))
;
set sashelp.class;
run;
data test;
set work.%LDsNm( peanut butter & jelly, a hot-dog in [a box], and s(*)t(*)a(*)r(*)s!! ) indsname=i;
indsname=symget(scan(i,-1,"."));
run;EXAMPLE 3.
data work.%LDsNm( . );
set sashelp.class;
run;
data %LDsNm( );
set sashelp.class;
run;
data %LDsNm();
set sashelp.class;
run;The LVarNm() macro function works like the LDSN() macro function, but for variables. Supported by LVarNmLab() macro function which allows to remember "user names" in labels.
The motivation for the macro was similar one as for the LDSN() macro.
See examples below for the details.
The %LVarNm() macro executes like a pure macro code.
Known "Limitations":
- variable name cannot contain unpaired quotes (
'and"), text:a "hot-dog"is ok, butJohn's dogis not!
Behaviour:
-
variable name text is converted to upcase
-
variable name text leading and trailing spaces are ignored, e.g. the following will give the same hash digest:
%LVarNm(test),%LVarNm( test),%LVarNm(test ). -
if the user want to add an extra suffix to the variable, e.g. to get a numerical suffix, the
%LVarNm()macro has to be wrapped inside the%unquote()macro function.
data test4;
array X[*] %unquote(%LVarNm(some strange! name))_0 - %unquote(%LVarNm(some strange! name))_10;
do i = lbound(X) to hbound(X);
X[i] = 2**(i-1);
put X[i]=;
end;
run;The reason for this is a "bug" like behaviour of SAS tokenizer/macroprocesor.
See the following SAS-L discussion thread:
https://listserv.uga.edu/scripts/wa-UGA.exe?A2=SAS-L;4b2bcf80.2205A&S=
The basic syntax is the following, the <...> means optional parameters:
%LVarNm(
arbitrary text string (in line with limitations)
)EXAMPLE 1.
options ls=max;
data test;
%LVarNmLab( peanut butter & jelly with a "Hot-Dog" in [a box], popcorn, and s*t*a*r*s )
do %LVarNm( peanut butter & jelly with a "Hot-Dog" in [a box], popcorn, and s*t*a*r*s ) = 1 to 10;
y = 5 + %LVarNm( peanut butter & jelly with a "Hot-Dog" in [a box], popcorn, and s*t*a*r*s ) * 17;
output;
end;
run;
data test2;
set test;
where %LVarNm( peanut butter & jelly with a "Hot-Dog" in [a box], popcorn, and s*t*a*r*s ) < 5;
run;EXAMPLE 2.
data test3;
%LVarNmLab() = 17;
%LVarNm() = 17;
%LVarNm( ) = 42;
%LVarNm( ) = 303;
run;EXAMPLE 3.
data test3;
%LVarNm(test) = 1;
%LVarNm( test) = 2;
%LVarNm(test ) = 3;
run;EXAMPLE 4.
data test4;
array X[*] %LVarNm(some strange! name)_0 - %LVarNm(some strange! name)_10;
do i = lbound(X) to hbound(X);
X[i] = 2**(i-1);
put X[i]=;
end;
run;The LVarNmLab() macro function supports LVarNm() and allows to remember "user names" in labels.
The motivation for the macro was similar one as for the LDSN() macro.
See examples in LVarNm() documentation for the details.
The %LVarNmLab() macro executes like a pure macro code.
Known "Limitations":
- variable name cannot contain unpaired quotes (
'and"), text:a "hot-dog"is ok, butJohn's dogis not!
Behaviour:
-
variable name text is converted to upcase
-
variable name text leading and trailing spaces are ignored, e.g. the following will give the same hash digest:
%LVarNmLab(test),%LVarNmLab( test),%LVarNmLab(test ).
The basic syntax is the following, the <...> means optional parameters:
%LVarNmLab(
arbitrary text string (in line with limitations)
)The %QdedupListX() macro deletes duplicated values from
a X-separated list of values, where the X represents
a single character separator. List, including separators,
can be no longer than a value carried by a single macro variable.
Caution. The value of X has to be in the first byte of the list,
just after the opening bracket, i.e. (X...).
Returned value is quoted with %superq(). Leading and trailing spaces are ignored.
The %QdedupListX() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%QdedupListX(
XlistXofXxXseparatedXvalues
)Arguments description:
list- A list of X-separated values.
EXAMPLE 1. Basic use-case one. Delete duplicated values from a list.
%put *%QdedupListX(|a|b|c|b|c)*;
%put *%QdedupListX( a b c b c)*;
%put *%QdedupListX(,a,b,c,b,c)*;
%put *%QdedupListX(XaXbXcXbXc)*;
%put *%QdedupListX(/a/b/c/b/c)*;
%put *%QdedupListX(%str(;a;b;c;b;c))*;
%put *%QdedupListX(%nrstr(&a&b&c&b&c))*;
%put *%QdedupListX(%nrstr(%a%b%c%b%c))*;EXAMPLE 2. Leading and trailing spaces are ignored. Delete duplicated values from a list.
%put *%QdedupListX(| a | b.b | c | b.b| c )*;
%put *%QdedupListX(. a . b b . c . b b. c )*;EXAMPLE 3. Macro variable as an argument. Delete duplicated values from a list.
%let list = 4$5.5$6$1$2$3$1$2$3$4$5.5$6;
%put *%QdedupListX($&list.)*;
%let list = 4$ 5.5$ 6$ 1$ 2$ 3$ 1$ 2$ 3$ 4$ 5.5$ 6$;
%put *%QdedupListX( &list.)*;The getVars() and QgetVars() macro functions allow to extract variables names form a dataset according to a given pattern into a list.
The getVars() returns unquoted value [by %unquote()]. The QgetVars() returns quoted value [by %superq()].
The %QgetVars() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%QgetVars(
ds
<,sep=>
<,pattern=>
<,varRange=>
<,quote=>
<,ignoreCases=>
)Arguments description:
ds- Required, the name of the dataset from which variables are to be taken.
-
sep = %str( )- Optional, default value%str( ), a variables separator on the created list. -
pattern = .*- Optional, default value.*(i.e. any text), a variable name regexp pattern, case INSENSITIVE! -
varRange = _all_- Optional, default value_all_, a named range list of variables. -
quote =- Optional, default value is blank, a quotation symbol to be used around values. -
ignoreCases=- Optional, default value is 1.
Indicates if search should be case insensitive.
See examples in %getVars() help for the details.
The zipEvalf() and QzipEvalf() macro functions allow to use a function on elements of pair of space-separated lists.
For two space-separated lists of text strings the corresponding elements are taken and the macro applies a function, provided by user, to calculate result of the function on taken elements.
When one of the lists is shorter then elements are "reused" starting from the beginning.
The zipEvalf() returns unquoted value [by %unquote()]. The QzipEvalf() returns quoted value [by %superq()].
See examples below for the details.
The %QzipEvalf() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%QzipEvalf(
first
,second
<,function=>
<,operator=>
<,argBf=>
<,argMd=>
<,argAf=>
<,format=>
)Arguments description:
-
first- Required, a space-separated list of texts. -
second- Required, a space-separated list of texts.
-
function = cat- Optional, default value iscat, a function which will be applied to corresponding pairs of elements of the first and the second list. -
operator =- Optional, default value is empty, arithmetic infix operator used with elements the first and the second list. The first list is used on the left side of the operator the second list is used on the right side of the operator. -
argBf =- Optional, default value is empty, arguments of the function inserted before elements the first list. If multiple should be comma-separated. -
argMd =- Optional, default value is empty, arguments of the function inserted between elements the first list and the second list. If multiple should be comma-separated. -
argAf =- Optional, default value is empty, arguments of the function inserted after elements the second list. If multiple should be comma-separated. -
format=- Optional, default value is empty, indicates a format which should be used to format the result, does not work when theoperator=is used.
See examples in %zipEvalf() help for the details.
The RainCloudPlot() macro allow to plot Rain Cloud plots, i.e. plots of kernel density estimates, jitter data values, and box-and-whiskers plot.
See examples below for the details.
The "Here Comes the Rain (Cloud Plot) Again" (B. Jablonski, 2024) article describing the macro is attached as an additional content to the package.
The basic syntax is the following, the <...> means optional parameters:
%RainCloudPlot(
DS
,gr
,vars
<,WidthPX=>
<,HeightPX=>
<,boxPlot=>
<,roundFactor=>
<,rainDropSize=>
<,boxPlotSymbolSize=>
<,boxPlotLineSize=>
<,boxPlotFill=>
<,whiskerScale=>
<,meanShiftLine=>
<,meanShiftStep=>
<,meanShiftColors=>
<,colorsList=>
<,monochrome=>
<,antialiasMax=>
<,title=>
<,footnote=>
<,catLabels=>
<,xLabels=>
<,xBothAxis=>
<,catLabelPos=>
<,xLabelPos=>
<,catLabelAttrs=>
<,xLabelAttrs=>
<,formated=>
<,y2axis=>
<,y2axisLevels=>
<,y2axisValueAttrs=>
<,y2axisFormat=>
<,y2axisLines=>
<,catAxisValueAttrs=>
<,xaxisValueAttrs=>
<,xaxisTickstyle=>
<,xaxisValues=>
<,xaxisValuesDisplay=>
<,xaxisValuesFormat=>
<,xaxisValuesRotate=>
<,xaxisOther=>
<,sganno=>
<,odsGraphicsOptions=>
<,sgPlotOptions=>
<,vertical=>
<,VSCALE=>
<,KERNEL_K=>
<,KERNEL_C=>
<,VSCALEmax=>
<,minRange=>
<,maxRange=>
<,cleanTempData=>
<,codePreview=>
<,reuseN=>
)Arguments description:
-
DS- Required, name of the dataset from which variables are to be taken. -
gr- Required, name of the grouping variable. When more than one variable is specified separate plots are rendered. Can be numeric or character. -
vars- Required, name of the aggregated numeric variable. When more than one variable is specified separate plots are rendered.
Plot related options:
-
WidthPX- Optional, default value1200. Total width of the plot in pixels. -
HeightPX- Optional, default value220. Partial height of the plot in pixels. Total height is calculated as#GROUPS x HeightPX. -
boxPlot- Optional, default value1. Indicates if the Box Plot should be added. -
roundFactor- Optional, default value0.000001. Rounding level when calculating maximum value of the cloud chart. Should be adjusted to data granularity level, e.g. for data with value around1e-8should be decreased. -
rainDropSize- Optional, default value5px. Size of data points in the "rain" plot. -
boxPlotSymbolSize- Optional, default value8px. Size of symbols on the box plot. If two values are provided, e.g.,16px 8px, the first is used for diamond size (the mean), the second for "min/max" bars. -
boxPlotLineSize- Optional, default value1px. Thickness of lines of the box plot. -
boxPlotFill- Optional, default value1. Transparency of the box plot. Ranges from 0.0 (opaque) to 1.0 (full translucent). -
whiskerScale=- Optional, default value1.5.
It providesWSparameter in theQ1 - WS*IQRformula and theQ3 + WS*IQRformula. Provided value should be a positive number. Otherwise it is set to 0. -
meanShiftLine- Optional, default value0. Indicates if a line connecting mean symbol on the Box Plot should be added. -
meanShiftStep- Optional, default value0.1. Sets how smooth gradient should be on the "mean-shift" line. -
meanShiftColors- Optional, default value is empty. List of colours for plotting the "mean-shift" line. Empty indicates that thecolorsListvalue will be used. -
colorsList- Optional, default value is empty. List of colours for plotting. Empty indicates that the default list will be used. -
monochrome- Optional, default value0. Indicates if the default list of colours should be gray-scale. -
antialiasMax- Optional, default value is empty. Sets a value to the ODS graphicsANTIALIASMAXoption. When empty the value is calculated from data. -
title- Optional, default value - see notes below. Provides a list of titles printed on the plot. For details see notes below. -
footnote- Optional, default value - see notes below. Provides a list of titles printed on the plot. For details see notes below. -
catLabels- Optional, default value is empty. List of values for group axis labels (vertical). When empty a grouping variable name is used. For details see notes below. -
xLabels- Optional, default value is empty. List of values for data variable axis labels (horizontal). When empty a data variable name is used. For details see notes below. -
xBothAxis- Optional, default value is1. Indicates if both (top and bottom) axis (horizontal) should be printed. If not1then only bottom axis is displayed. -
catLabelPos- Optional, default valueDATACENTER. Indicates position of the label on group axis (vertical). Allowed values areBOTTOM,CENTER,DATACENTER, andTOP. -
xLabelPos- Optional, default valueDATACENTER. Indicates position of the label on data axis (horizontal). Allowed values areLEFT,CENTER,DATACENTER, andRIGHT. -
catLabelAttrs- Optional, default value is empty. List of attributes for group axis labels (vertical). For details see notes below. -
xLabelAttrs- Optional, default value is empty. List of attributes for data variable axis labels (horizontal). For details see notes below. -
formated- Optional, default value0. Indicates if values of the grouping variable should be formatted. -
y2axis- Optional, default value1. Indicates if the right vertical axis should be displayed. -
y2axisLevels- Optional, default value4. Sets the number of expected levels of values printed on the right vertical axis. -
y2axisValueAttrs- Optional, default valueColor=Grey. Allows to modify Y2 axis values attributes. -
y2axisFormat- Optional, default value12.2-L. Allows to modify Y2 axis values format. -
y2axisLines- Optional, default value0. If1, adds horizontal lines for Y2 axis values. -
catAxisValueAttrs- Optional, default valueColor=Black. Allows to modify category (Y) axis values attributes. -
xaxisValueAttrs- Optional, default valueColor=Grey. Allows to modify X axis values attributes. -
xaxisTickstyle- Optional, default valueINSIDE. Allows to modify X axis tick style. Allowed values areOUTSIDE,INSIDE,ACROSS, andINBETWEEN. For SAS previous to 9.4M5 set to missing! -
xaxisValues- Optional, default value is empty. It is a wrapper to provide value for theXAXISstatement, forValuesoption. -
xaxisValuesDisplay- Optional, default value is empty. It is a wrapper to provide value for theXAXISstatement, forValuesDisplayoption. -
xaxisValuesFormat- Optional, default value is empty. It is a wrapper to provide value for theXAXISstatement, forValuesFormatoption. Instead usingw.dformat, use its aliasFw.d. -
xaxisValuesRotate- Optional, default value is empty. It is a wrapper to provide value for theXAXISstatement, forValuesRotateoption. -
xaxisOther- Optional, default value is empty. It is a wrapper to provide value for theXAXISstatement options not mentioned above. You can use it to provide, e.g.,GRID, 'MINOR',MINORGRID,GRIDATTRS=, orMINORGRIDATTRS=. -
sganno- Optional, default value is empty. keeps name of a data set for thesganno=option of the SGPLOT procedure. -
sgPlotOptions- Optional, default value isnoautolegend noborder subpixel. List of additional options values for SGPLOT procedure. -
odsGraphicsOptions- Optional, default value is empty. List of additional options values forODS Graphicsstatement. By default only the:width=,height=, andantialiasmax=are modified. -
vertical- Optional, default value is0. Set value to1to plot "clouds & boxes" vertically. NOTE: Before setting the parameter to1, first prepare the plot in the "horizontal" version since all other parameters assume that orientation(!) and then are converted accordingly.
Stat related options:
-
VSCALE- Optional, default valueProportion. Specifies the scale of the vertical axis. Allowed values arePROPORTION,PERCENT, andCOUNT.PROPORTIONscales the data in units of proportion of observations per data unit.PERCENTscales the data in units of percent of observations per data unit.COUNTscales the data in units of the number of observations per data unit. -
KERNEL_K- Optional, default valueNORMAL. Specifies type of kernel function to compute kernel density estimates. Allowed values areNORMAL,QUADRATIC, andTRIANGULAR. -
KERNEL_C- Optional, default value1. Specifies standardized bandwidth parameter C to compute kernel density estimates. Allowed values are between0and1, -
VSCALEmax- Optional, default value is empty. Provides maximal value for Y2 axis scale. When used an information note is displayed in the log. Recommended values are: ForVSCALE=PERCENTbetween 0 and 100, forVSCALE=PROPORTIONbetween 0 and 1, and forVSCALE=COUNTbetween 0 and N (sample size). -
minRange- Optional, default value is.(numerical missing). Indicates minimum value for x-axis on the plot, by default calculated form data. Is a global parameter used for all plots. -
maxRange- Optional, default value is.(numerical missing). Indicates maximum value for x-axis on the plot, by default calculated form data. Is a global parameter used for all plots.
Other options:
-
cleanTempData- Optional, default value1. Indicates if temporary data sets should be deleted. -
codePreview- Optional, default value0. Indicates if source code should be MPRINTed to log. -
reuseN- Optional, default value6. Indicates how many times colours list should be repeated so that colours could be reused in case the number of groups is greater than the colours list length.
-
Default value of the
titleoption is:%nrstr(title1 JUSTIFY=C "Rain Cloud plot for &list_g. by " %unquote(&xLabel.);)Use the%str()or%nrstr()macro-function to handle special characters. The%unquote()is used when resolving the parameter. -
Default value of the
footnoteoption is:%nrstr(footnote1 JUSTIFY=L COLOR=lightGray HEIGHT=1 "by RainCloudPlot macro from the BasePlus package";)Use the%str()or%nrstr()macro-function to handle special characters. The%unquote()is used when resolving the parameter. -
The
catLabelsandxLabelsshould be quoted comma-separated lists enclosed with brackets, e.g.catLabels=("Continent of Origin", "Car Type"), see Example below. -
The
catLabelAttrsandxLabelAttrsshould be space-separated lists ofkey=valuepairs, e.g.xLabelAttrs=size=12 color=Pink weight=bold, see Example below. -
Kernel density estimates and basic statistics are calculated with
PROC UNIVARIATE. -
Plot is generated by
PROC SGPLOTwithBAND,SCATTER, andPOLYGONplots. -
After execution the ODS graphics dimension parameters are set to
800pxby600px. -
SAS notes (
NOTE:) are disabled for the execution time. -
Before setting the
vertical=parameter to1, first prepare the plot in the "horizontal" version since all other parameters assume that orientation(!) and then are converted accordingly. -
List of predefined colours is:
BlueViolet,RoyalBlue,OliveDrab,Gold,HotPink,Crimson,
MediumPurple,CornflowerBlue,YellowGreen,Goldenrod,Orchid,IndianRed.
The box-and-whiskers plot has the following interpretation:
- left vertical bar indicates the minimum,
- left whisker line starts at
max(Q1 - WS*IQR, minimum)and ends at lower quartile (Q1), - diamond indicates mean,
- vertical bar inside of the box indicates median,
- right whisker line starts at upper quartile (Q3) and ends at
min(Q3 + WS*IQR, maximum), - right vertical bar indicates the maximum.
The WS value is provided through whiskerScale= parameter. Default value is 1.5.
With above setup it may happen that there is a gap between the minimum marker and the beginning of the left whisker or there is a gap between the end of the right whisker and the maximum marker. See examples below.
EXAMPLE 1. Simple Rain Cloud Plot for a have dataset:
data have;
g = "Aaa";
do _N_ = 1 to 50;
x = rannor(42);
y = ranuni(42);
output;
end;
g = "Bb";
do _N_ = 1 to 120;
select (mod(_N_,9));
when(1,2,3,4,5) x = 0.5*rannor(42)+1;
when(6,7,8) x = 0.5*rannor(42)+3;
otherwise x = 0.5*rannor(42)+5;
end;
y = ranuni(42)+1;
output;
end;
g = "C";
do _N_ = 1 to 60;
x = 3*rannor(42)+7;
y = ranuni(42)+2;
output;
end;
run;
%RainCloudPlot(have, g, x y)
The output can be seen in the md file.
EXAMPLE 2. Rain Cloud plot for sashelp.cars dataset
with groups by Origin or Type
for Invoice variables:
%RainCloudPlot(
sashelp.cars(where=(Type ne "Hybrid"))
, Origin Type
, Invoice
, HeightPX=300
, y2axisLevels=3
, catLabels=("Continent of Origin", "Car Type")
, xLabels="Invoice, [$]"
, xLabelAttrs=size=12 color=Pink weight=bold
)
The output can be seen in the md file.
EXAMPLE 3. Rain Cloud plot with formated groups: and annotations.
data annotation;
function="text";
label="This graph is full(*ESC*){sup '2'} of annotations!";
drawspace="graphpercent";
rotate=30;
anchor="center";
textsize=32;
x1=50;
y1=50;
textcolor="red";
justify="center";
textweight="bold";
width=100;
widthunit="percent";
run;
proc format;
value system
1="Windows"
2="MacOS"
3="Linux"
;
run;
data test;
do system = 1 to 3;
do i = 1 to 50;
x = 2*rannor(123)/system - system;
output;
end;
end;
format system system.;
run;
%RainCloudPlot(test, system, x
, colorslist=CX88CCEE CX44AA99 CX117733
, formated=1
, sganno=annotation
, sgPlotOptions=noborder
, WidthPX=1000
, HeightPX=380
, catAxisValueAttrs=Color=Green weight=bold
, VSCALE=percent
, cleanTempData=0
, y2axisLevels=5
, y2axisFormat=words20.-L
, VSCALEmax=60
, codePreview=1
, y2axisValueAttrs=Color=Grey size=10px
, y2axisLines=1
, xLabels="variable X"
, title = %nrstr(title1;
title2 JUSTIFY=L "Rain Cloud plot for &list_g. by " %unquote(&xLabel.);
title3 JUSTIFY=C "Rain Cloud plot for &list_g. by " %unquote(&xLabel.);
title4 JUSTIFY=R "Rain Cloud plot for &list_g. by " %unquote(&xLabel.);
)
)
The output can be seen in the md file.
EXAMPLE 4. Rain Cloud plot for sashelp.cars dataset
with groups by Drive Train for Weight (LBS)
variable ploted "vertically":
%RainCloudPlot(
sashelp.cars
, DriveTrain
, Weight
, HeightPX=400
, colorslist=Red Green Blue
, y2axisLevels=5
, catLabels=("DriveTrain")
, xLabels="Weight (LBS)"
, xLabelAttrs=size=12 color=Black weight=bold
, y2axisLines=1
, vscale=percent
, vscalemax=50
, vertical = 1
, title = %nrstr(title1 J=C HEIGHT=3 "The VERTICAL plotting is cool, ...";)
, footnote = %nrstr(footnote1 J=L HEIGHT=2 "... isn't it?";)
, whiskerScale = 1.5
)
The output can be seen in the md file.
The repeatTxt() macro function allows to repeat n
times a text string separated by string s=.
The repeatTxt() returns unquoted value [by %unquote()].
See examples below for the details.
The %repeatTxt() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%repeatTxt(
text
<,n>
<,s=>
)Arguments description:
-
text- Required, a text to be repeated. -
n- Required/Optional, the number of repetitions. If missing then set to1;
s = %str( )- Optional, it is a separator between repeated elements. Default value is space.
EXAMPLE 1. Simple repetition of dataset name:
options mprint;
data work.test5;
set
%repeatTxt(sashelp.cars, 5)
;
run;EXAMPLE 2. Simple repetition of data step:
options mprint;
%repeatTxt(data _null_; set sashelp.cars; run;, 3)EXAMPLE 3. "Nice" output:
%put %repeatTxt(#,15,s=$) HELLO SAS! %repeatTxt(#,15,s=$);EXAMPLE 4. Macroquote a text with commas:
%repeatTxt(
%str(proc sql; create table wh as select weight,height from sashelp.class; quit;)
,3
) EXAMPLE 5. Empty n repeats text one time:
options mprint;
data work.test1;
set
%repeatTxt(sashelp.cars)
;
run;EXAMPLE 6. Dynamic "formatting":
%macro printWork();
%let work=%sysfunc(pathname(work));
%put +%repeatTxt(~,%length(&work.)+5,s=)+;
%put {&=work.};
%put +%repeatTxt(~,%length(&work.)+5,s=)+;
%mend printWork;
%printWork()The splitDSIntoBlocks() macro allows to split the set dataset into blocks
of size blockSize in datasets: prefix1 to prefixN.
The last dataset may have less observations then the blockSize.
Macro covers BASE engine (v9, v8, v7, v6) and SPDE engine datasets.
See examples below for the details.
The basic syntax is the following, the <...> means optional parameters:
%splitDSIntoBlocks(
blockSize
<,set>
<,prefix>
)Arguments description:
-
blockSize- Required, the size of the block of data, in other words number of observations in one block of split data. Block size must be positive integer. -
set- Required/Optional, the name of the dataset to split. If empty then&syslast.is used. -
prefix- Required/Optional, the name-prefix for new datasets. If missing then set topart.
EXAMPLE 1. Split sashelp.class into 5 elements datasets ABC1 to ABC4:
%splitDSIntoBlocks(5,sashelp.class,ABC)EXAMPLE 2. By default splits the _last_ dataset into part1 to partN datasets:
data lastData;
set sashelp.cars;
run;
%splitDSIntoBlocks(123)EXAMPLE 3. Works with SPDE engine too:
options dlcreatedir;
libname test "%sysfunc(pathname(work))/testSPDE";
libname test;
libname test SPDE "%sysfunc(pathname(work))/testSPDE";
data test.test;
set sashelp.cars;
run;
%splitDSIntoBlocks(100,test.test,work.spde)The splitDSIntoParts() macro allows to split the set dataset into parts parts
of approximately NOBS/parts size in datasets: prefix1 to prefixN.
The splitDSIntoParts() macro internally runs the splitDSIntoBlocks() macro.
Macro covers BASE engine (v9, v8, v7, v6) and SPDE engine datasets.
See examples below for the details.
The basic syntax is the following, the <...> means optional parameters:
%splitDSIntoParts(
parts
<,set>
<,prefix>
)Arguments description:
-
parts- Required, the number of parts to split data into. Number of parts must be positive integer. -
set- Required/Optional, the name of the dataset to split. If empty then&syslast.is used. -
prefix- Required/Optional, the name-prefix for new datasets. If missing then set topart.
EXAMPLE 1. Split sashelp.cars into 7 parts: datasets carsInParts1 to carsInParts7:
%splitDSIntoParts(7,sashelp.cars, carsInParts)EXAMPLE 2. By default splits the _last_ dataset into part1 to part3 datasets:
data lastData;
set sashelp.cars;
run;
%splitDSIntoBlocks(3)EXAMPLE 3. Works with SPDE engine too:
options dlcreatedir;
libname test "%sysfunc(pathname(work))/testSPDE";
libname test;
libname test SPDE "%sysfunc(pathname(work))/testSPDE";
data test.test;
set sashelp.cars;
run;
%splitDSIntoParts(3,test.test,work.spde)The %symdelGlobal() macro deletes all global macro variables
created by the user. The only exceptions are read only variables
and variables the one which starts with SYS, AF, or FSP.
In that case a warning is printed in the log.
One temporary global macro variable ________________98_76_54_32_10_
and a dataset, in work library, named _%sysfunc(datetime(),hex7.)
are created and deleted during the process.
The %symdelGlobal() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%symdelGlobal(
info
)Arguments description:
info- Optional, default value should be empty, if set toNOINFOorQUIETthen infos and warnings about variables deletion are suspended.
EXAMPLE 1. Basic use-case one. Delete global macro variables, info notes and warnings are printed in the log.
%let a = 1;
%let b = 2;
%let c = 3;
%let sys_my_var = 11;
%let af_my_var = 22;
%let fsp_my_var = 33;
%global / readonly read_only_x = 1234567890;
%put _user_;
%symdelGlobal();
%put _user_;EXAMPLE 2. Basic use-case two. Delete global macro variables in quite mode No info notes and warnings are printed in the log.
%let a = 1;
%let b = 2;
%let c = 3;
%let sys_my_var = 11;
%let af_my_var = 22;
%let fsp_my_var = 33;
%global / readonly read_only_x = 1234567890;
%put _user_;
%put *%symdelGlobal(NOINFO)*;
%put _user_;The unzipArch() macro allows to unzip content of a ZIP archive.
Macro is OS-independent, the XCMD option is not required.
The dlCreateDir option is used under the hood.
Content of unzipped archive can be listed in the log.
Source files can be deleted after decompression. Errors of decompression and are reported. If any occur the deletion is suspended.
See examples below for the details.
The basic syntax is the following, the <...> means optional parameters:
%unzipArch(
archName
<,path=>
<,pathRef=>
<,target=>
<,targetRef=>
<,list=>
<,clean=>
)Arguments description:
archName- Required, name of the ZIP archive to be extracted. Name should be full, i.e., with the extension!
-
path=- Optional, a path pointing to zipped file location. The path should be provided unquoted. Default value isWORKlocation. -
pathRef=- Optional, a fileref to path pointing to zipped file location. Thepath, if not null, has precedense over thepathRef. -
target=- Optional, a path pointing to target location where files will be extracted. The path should be provided unquoted. Default value isWORKlocation. -
target=- Optional, a fileref to path pointing to target location where files will be extracted. Thetarget, if not null, has precedense over thetargetRef. -
list = 0- Optional, default value is0, indicates if zip content should be listed in the log.1means yes,0means no. -
clean = 0- Optional, default value is0, indicates if zip file should be deleted after unzipping.1means yes,0means no.
EXAMPLE 1. Unzip compressed archive. Example requires the basePlus package.
filename arch ZIP "%workPath()/testArch.zip";
data _null_;
file arch(abc/test1.txt);
put "text for test file 1";
data _null_;
file arch(abc/subdir/test2.txt);
put "text for test file 2";
data _null_;
file arch(abc/subdir/test3.txt);
put "text for test file 3";
run;
%unzipArch(
testArch.zip
, path = %workPath()
, target = %workPath()
, list=1
);
filename pR "%workPath()";
%unzipArch(
testArch.zip
, pathRef = pR
, targetRef = pR
, clean=1
);
The unzipLibrary() macro allows to unzip content of a SAS library.
It is a counterpart to the %zipLibrary() macro and is not intended to work
with zip files generated by other software (though it may in some cases).
Files can be unzipped from a single file or from multiple files (named e.g. "dataset.sas7bdat.zip"). If a file is indexed also the index file is unzipped.
Source files can be deleted after decompression.
Status of decompression and processing time is reported.
See examples below for the details.
The basic syntax is the following, the <...> means optional parameters:
%unzipLibrary(
path
<,zip=>
<,mode=>
<,clean=>
<,libOut=>
)Arguments description:
path- Required, a path pointing to zipped file(s) location. The path should be unquoted, e.g.%unzipLibrary(/some/dir, ...).
-
zip =- Optional, Whenmode=Sa name of the zip file containing SAS files to be unzipped. -
mode = S- Optional, default value isS, indicates mode of decompression read from a single zip file (SINGLE/S) or from multiple files (MULTI/M) -
clean = 0- Optional, default value is0, should zip files be deleted after unzipping?1means yes,0means no. -
libOut =- Optional, default value is empty, output library for a single zip file decompression.
EXAMPLE 1. Generate data:
options dlcreatedir;
libname test1 "%sysfunc(pathname(work))/test1";
libname test2 "%sysfunc(pathname(work))/test2";
libname test3 (test1 test2);
libname test4 "%sysfunc(pathname(work))/test4";
options nodlcreatedir;
%put %sysfunc(pathname(test3));
%put %sysfunc(pathname(test4));
data
test1.A(index=(model))
test1.B
test2.C
test2.D(index=(model make io=(invoice origin)))
;
set sashelp.cars;
run;
data test1.B2 / view=test1.B2;
set test1.B;
output;
output;
run;EXAMPLE 2. Use data from Example 1.
First zip content of the test3 library
to test4 location into one zip file
and delete source files.
Next unzip test3.zip library into the
test4 location and delete the zip file.
%zipLibrary(test3, clean=1, libOut=test4)
%unzipLibrary(%sysfunc(pathname(test4)), zip=test3, clean=1)EXAMPLE 3. Use data from Example 1.
First zip content of the test1 library
into multiple zip files and delete source files.
Next unzip *.zip files in test1
location and delete zipped files
%zipLibrary(test1, mode=M, clean=1)
%unzipLibrary(%sysfunc(pathname(test1)), mode=M, clean=1)EXAMPLE 4. First zip content of the sashelp library
into work library.
Next unzip sashelp.zip file in work
location and delete zip file.
%zipLibrary(sashelp, mode=S, clean=0, libOut=work)
%unzipLibrary(%sysfunc(pathname(work)), zip=sashelp, mode=S, clean=1)The zipArch() macro allows to ZIP content of a directory.
Macro is OS-independent, the XCMD option is not required.
Content of zipped archive can be listed in the log.
Errors of decompression and are reported.
Macro does not include hidden files.
See examples below for the details.
The basic syntax is the following, the <...> means optional parameters:
%zipArch(
archName
,path =
<,pathRef=>
<,target=>
<,targetRef=>
<,list=>
<,overwrite=>
<,dropList=>
)Arguments description:
-
archName- Required, name of the archive to be generated. Name should be full, i.e., with the extension! -
path=- Required/Optional, location of a directory to ZIP. The path should be provided unquoted. Has priority over thepathRefparameter.
-
pathRef=- Required/Optional, fileref to location of a directory to ZIP. The reference has to be pointing to single directory. If provided withpath- thepathtakes priority over thepathRefparameter. -
target=- Optional, a path pointing to target location where the archive will be generated. The path should be provided unquoted. Default value isWORKlocation. Has priority over thetargetRefparameter. -
targetRef=- Optional, fileref to a path pointing to target location where the archive will be generated. The reference has to be pointing to single directory. If provided withtarget- thetargettakes priority over thetargetRefparameter. -
list = 0- Optional, default value is0, indicates if zip content should be listed in the log.1means yes,0means no. -
overwrite = 0- Optional, default value is0, indicates if existing archive file should be overwritten.1means yes,0means no. -
overwrite = 1- Technical, default value is1, indicates if the "to-be-zipped-files-list" data set should be deleted.1means yes,0means no.
EXAMPLE 1. Zip a directory . Example requires the basePlus package.
options dlCreateDir;
libname arch1 "%workPath()/testArch1";
libname arch2 "%workPath()/testArch2";
filename arch1 "%workPath()/testArch1";
data _null_;
file arch1(test1.txt);
put "text for test file 1";
data _null_;
file arch1(test2.txt);
put "text for test file 2";
data _null_;
file arch1(test3.txt);
put "text for test file 3";
run;
data arch1.class(index=(name));
set sashelp.class;
run;
data arch1.cars(index=(model));
set sashelp.cars;
run;
%zipArch(
archName1.zip
, path = %workPath()/testArch1
, list = 1
, overwrite = 1
)
%zipArch(
archName2.zip
, pathRef = arch1
, target = %workPath()/testArch2
, list = 1
, overwrite = 1
)
%unzipArch(
archName2.zip
, path = %workPath()/testArch2
, target = %workPath()/testArch2
, clean=1
, list=1
);
The zipEvalf() and QzipEvalf() macro functions allow to use a function on elements of pair of space-separated lists.
For two space-separated lists of text strings the corresponding elements are taken and the macro applies a function, provided by user, to calculate result of the function on taken elements.
When one of the lists is shorter then elements are "reused" starting from the beginning.
The zipEvalf() returns unquoted value [by %unquote()]. The QzipEvalf() returns quoted value [by %superq()].
See examples below for the details.
The %zipEvalf() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%zipEvalf(
first
,second
<,function=>
<,operator=>
<,argBf=>
<,argMd=>
<,argAf=>
<,format=>
)Arguments description:
-
first- Required, a space-separated list of texts. -
second- Required, a space-separated list of texts.
-
function = cat- Optional, default value iscat, a function which will be applied to corresponding pairs of elements of the first and the second list. -
operator =- Optional, default value is empty, arithmetic infix operator used with elements the first and the second list. The first list is used on the left side of the operator the second list is used on the right side of the operator. -
argBf =- Optional, default value is empty, arguments of the function inserted before elements the first list. If multiple should be comma-separated. -
argMd =- Optional, default value is empty, arguments of the function inserted between elements the first list and the second list. If multiple should be comma-separated. -
argAf =- Optional, default value is empty, arguments of the function inserted after elements the second list. If multiple should be comma-separated. -
format=- Optional, default value is empty, indicates a format which should be used to format the result, does not work when theoperator=is used.
EXAMPLE 1. Simple concatenation of elements:
%let x = %zipEvalf(1 2 3 4 5 6, q w e r t y);
%put &=x;EXAMPLE 2. Shorter list is "reused":
%let x = %zipEvalf(1 2 3 4 5 6, a b c);
%put &=x;EXAMPLE 3. Use of the operator=, shorter list is "reused":
%let y = %zipEvalf(1 2 3 4 5 6, 100 200, operator = +);
%put &=y;
%let z = %zipEvalf(1 2 3 4 5 6 8 9 10, 1 2 3 4 5 6 8 9 10, operator = **);
%put &=z;EXAMPLE 4. Format result:
%let x = %zipEvalf(1 2 3 4 5 6, q w e r t y, format=$upcase.);
%put &=x;
%put *
%zipEvalf(
ą ż ś ź ę ć ń ó ł
,Ą Ż Ś Ź Ę Ć Ń Ó Ł
,format = $brackets.
)
*;EXAMPLE 5. Use with macro variables:
%let abc = 10 100 1000;
%put *
%zipEvalf(
%str(1 2 3 4 5 6 7 8 9)
,&abc.
,function = sum
)
*;EXAMPLE 6. If one of elements is empty:
%put *
%zipEvalf(
abc efg
,
)
*;EXAMPLE 7. Use of the function=, shorter list is "reused":
%put *
%zipEvalf(
a b c
,efg
,function = catx
,argBf = %str(,)
,format = $brackets.
)
*;
%put *
%zipEvalf(
a b c
,efg
,function = catx
,argBf = %str( )
,format = $upcase.
)
*;
%put *
%zipEvalf(
%str(! @ # $ [ ] % ^ & * )
,1 2 3 4 5 6 7 8 9
,function = catx
,argBf = %str( )
,format = $quote.
)
*;EXAMPLE 8. Use inside resolve:
data _null_;
z = resolve('
%zipEvalf(
%nrstr(! @ # $ [ ] % ^ & *)
,1 2 3 4 5 6 7 8 9
,function = catx
,argBf = %str(.)
,format = $quote.
)');
put z=;
run;EXAMPLE 9. Use in data step:
data test;
%zipEvalf(
a b c d e f g
,1 2 3 4 5 6 7
,function = catx
,argBf = =
,format = $semicolon.
)
run;EXAMPLE 10. With 9.4M6 hashing() function:
%put %zipEvalf(MD5 SHA1 SHA256 SHA384 SHA512 CRC32, abcd, function = HASHING);EXAMPLE 11. Use middle argument:
%let x = %zipEvalf(1 2 3 4 5 6, 2020, argMd=5, function=MDY, format=date11.);
%put &=x;The zipLibrary() macro allows to zip content of a SAS library.
Files can be zipped into a single file (named as the input library) or into multiple files (named as "dataset.sas7bdat.zip"). If a file is indexed also the index file is zipped.
Source files can be deleted after compression.
Status of compression and processing time is reported.
See examples below for the details.
The basic syntax is the following, the <...> means optional parameters:
%zipLibrary(
lib
<,mode=>
<,clean=>
<,libOut=>
<,compression=>
)Arguments description:
lib- Required, a name of the library to be zipped. Must be a valid SAS V7, V8, or V9 library.
-
mode = S- Optional, default value isS, indicates mode of compression generates single zip file (SINGLE/S) or multiple files (MULTI/M) -
clean = 0- Optional, default value is0, should datasets be deleted after zipping?1means yes,0means no. -
libOut =- Optional, default value is empty, output library for a single zip file. -
compression =- Optional, default value is6, specifies the compression level0to9, where0is no compression and9is maximum compression.
EXAMPLE 1. Generate data:
options dlcreatedir;
libname test1 "%sysfunc(pathname(work))/test1";
libname test2 "%sysfunc(pathname(work))/test2";
libname test3 (test1 test2);
libname test4 "%sysfunc(pathname(work))/test4";
options nodlcreatedir;
%put %sysfunc(pathname(test3));
%put %sysfunc(pathname(test4));
data
test1.A(index=(model))
test1.B
test2.C
test2.D(index=(model make io=(invoice origin)))
;
set sashelp.cars;
run;
data test1.B2 / view=test1.B2;
set test1.B;
output;
output;
run;EXAMPLE 2. Zip content of test3 library into the same location in one zip file:
%zipLibrary(test3)EXAMPLE 3. Zip content of test3 library into the same location in multiple zip files:
%zipLibrary(test3, mode=MULTI)EXAMPLE 4. Zip content of test3 library with maximum compression level into different location in one zip file and delete source files:
%zipLibrary(test3, clean=1, libOut=test4, compression=9)The bool format returns: zero for 0 or missing, one for other values.
It allows for a %sysevalf()'ish
conversion-type [i.e. %sysevalf(1.7 & 4.2, boolean)]
inside the %sysfunc() [e.g. %sysfunc(aFunction(), bool.)]
The boolz format returns: zero for 0 or missing, one for other values.
Fuzz value is 0.
It allows for a %sysevalf()'ish
conversion-type [i.e. %sysevalf(1.7 & 4.2, boolean)]
inside the %sysfunc() [e.g. %sysfunc(aFunction(), boolz.)]
The ceil format is a "wrapper" for the ceil() function.
It allows for a %sysevalf()'ish
conversion-type [i.e. %sysevalf(1.7 + 4.2, ceil)]
inside the %sysfunc() [e.g. %sysfunc(aFunction(), ceil.)]
The floor format is a "wrapper" for the floor() function.
It allows for a %sysevalf()'ish
conversion-type [i.e. %sysevalf(1.7 + 4.2, floor)]
inside the %sysfunc() [e.g. %sysfunc(aFunction(), floor.)]
The int format is a "wrapper" for the int() function.
It allows for a %sysevalf()'ish
conversion-type [i.e. %sysevalf(1.7 + 4.2, integer)]
inside the %sysfunc() [e.g. %sysfunc(aFunction(), int.)]
The arrFill() subroutine is a wrapper for the Call Fillmatrix() [a special FCMP subroutine].
A numeric array is filled with selected numeric value, e.g.
for array A = [. . . .] the subroutine
call arrFill(42, A) returns A = [42 42 42 42]
Caution! Array parameters to subroutine calls must be 1-based.
The basic syntax is the following, the <...> means optional parameters:
call arrFill(N ,A)Arguments description:
-
N- Numeric value. -
A- Numeric array.
Example 1.
data _null_;
array X[*] a b c;
put "before: " (_all_) (=);
call arrFill(42, X);
put "after: " (_all_) (=);
run;The arrFillC() subroutine fills a character array with selected character value, e.g.
for array A = [" ", " ", " "] the subroutine
call arrFillC("B", A) returns A = ["B", "B", "B"]
Caution! Array parameters to subroutine calls must be 1-based.
The basic syntax is the following, the <...> means optional parameters:
call arrFillC(C ,A)Arguments description:
-
C- Character value. -
A- Character array.
Example 1.
data _null_;
array X[*] $ a b c;
put "before: " (_all_) (=);
call arrFillC("ABC", X);
put "after: " (_all_) (=);
run;The arrMissFill() subroutine fills
all missing values (i.e. less or equal than .Z)
of a numeric array with selected numeric value, e.g.
for array A = [1 . . 4] the subroutine
call arrMissFill(42, A) returns A = [1 42 42 4]
Caution! Array parameters to subroutine calls must be 1-based.
The basic syntax is the following, the <...> means optional parameters:
call arrMissFill(N ,A)Arguments description:
-
N- Numeric value. -
A- Numeric array.
Example 1.
data have;
input a b c;
cards4;
1 . 3
. 2 .
. . 3
;;;;
run;
data _null_;
set have ;
array X[*] a b c;
put "before: " (_all_) (=);
call arrMissFill(42, X);
put "after: " (_all_) (=);
run;The arrMissFillC() subroutine fills all missing values of a character array with selected character value, e.g.
for array A = ["A", " ", "C"] the subroutine
call arrMissFillC("B", A) returns A = ["A", "B", "C"]
Caution! Array parameters to subroutine calls must be 1-based.
The basic syntax is the following, the <...> means optional parameters:
call arrMissFillC(C, A)Arguments description:
-
C- Character value. -
A- Character array.
Example 1.
data have;
infile cards dsd dlm="," missover;
input (a b c) (: $ 1.);
cards4;
A, ,C
,B,
, ,C
;;;;
run;
data _null_;
set have ;
array X[*] $ a b c;
put "before: " (_all_) (=);
call arrMissFillC("X", X);
put "after: " (_all_) (=);
run;The arrMissToLeft() subroutine shifts
all non-missing (i.e. greater than .Z)
numeric elements to the right side of an array
and missing values to the left, e.g.
for array A = [1 . 2 . 3] the subroutine
call arrMissToLeft(A) returns A = [. . 1 2 3]
All missing values are replaced with the dot (.)
Caution! Array parameters to subroutine calls must be 1-based.
The basic syntax is the following, the <...> means optional parameters:
call arrMissToLeft(A)Arguments description:
A- Numeric array.
Example 1.
data have;
input a b c;
cards4;
1 . 3
. 2 .
. . 3
;;;;
run;
data _null_;
set have ;
array X[*] a b c;
put "before: " (_all_) (=);
call arrMissToLeft(X);
put "after: " (_all_) (=);
run;The arrMissToLeftC() subroutine shifts all non-missing (i.e. different than empty string) character elements to the right side of an array and all missing values to the left, e.g.
for array A = ["A", " ", "B", " ", "C"] the subroutine
call arrMissToLeftC(A) returns A = [" ", " ", "A", "B", "C"]
Caution! Array parameters to subroutine calls must be 1-based.
The basic syntax is the following, the <...> means optional parameters:
call arrMissToLeftC(A)Arguments description:
A- Character array.
Example 1.
data have;
infile cards dsd dlm="," missover;
input (a b c) (: $ 1.);
cards4;
A, ,C
,B,
, ,C
;;;;
run;
data _null_;
set have ;
array X[*] $ a b c;
put "before: " (_all_) (=);
call arrMissToLeftC(X);
put "after: " (_all_) (=);
run;The arrMissToRight() subroutine shifts
all non-missing (i.e. greater than .Z)
numeric elements to the left side of an array
and missing values to the right, e.g.
for array A = [1 . 2 . 3] the subroutine
call arrMissToRight(A) returns A = [1 2 3 . .]
All missing values are replaced with the dot (.)
Caution! Array parameters to subroutine calls must be 1-based.
The basic syntax is the following, the <...> means optional parameters:
call arrMissToRight(A)Arguments description:
A- Numeric array.
Example 1.
data have;
input a b c;
cards4;
1 . 3
. 2 .
. . 3
;;;;
run;
data _null_;
set have ;
array X[*] a b c;
put "before: " (_all_) (=);
call arrMissToRight(X);
put "after: " (_all_) (=);
run;The arrMissToRightC() subroutine shifts all non-missing (i.e. different than empty string) character elements to the left side of an array and missing values to the right, e.g.
for array A = ["A", " ", "B", " ", "C"] the subroutine
call arrMissToRightC(A) returns A = ["A", "B", "C", " ", " "]
Caution! Array parameters to subroutine calls must be 1-based.
The basic syntax is the following, the <...> means optional parameters:
call arrMissToRightC(A)Arguments description:
A- Character array.
Example 1.
data have;
infile cards dsd dlm="," missover;
input (a b c) (: $ 1.);
cards4;
A, ,C
,B,
, ,C
;;;;
run;
data _null_;
set have ;
array X[*] $ a b c;
put "before: " (_all_) (=);
call arrMissToRightC(X);
put "after: " (_all_) (=);
run;The bracketsC() function is internal function used by the brackets format. Returns character value of length 32767.
The basic syntax is the following:
bracketsC(X)Arguments description:
X- Character value.
The bracketsN() function is internal function used by the brackets format. Returns character value of length 34.
The basic syntax is the following:
bracketsN(X)Arguments description:
X- Numeric value.
The catXFc() function is a wrapper
of the catX() function but with ability
to format character values.
For array A = ["a", " ", "c"] the
catXFc("upcase.", "*", A) returns "A*C".
If format does not handle nulls they are ignored.
Caution! Array parameters to function calls must be 1-based.
The basic syntax is the following, the <...> means optional parameters:
catXFc(format, delimiter, A)Arguments description:
-
format- A name of the character format to be used. -
delimiter- A delimiter string to be used. -
A- Character array
Example 1.
data _null_;
t = "t";
u = " ";
v = "v";
array b[*] t u v;
length s $ 17;
s = catXFc("upcase.", "*", B);
put (_all_) (=);
run;The catXFi() function is a wrapper
of the catX() function but with ability
to format numeric values but
IGNORES missing values (i.e. ._, ., .a, ..., .z).
For array A = [0, ., 2] the
catXFi("date9.", "#", A) returns
"01JAN1960#03JAN1960"
Caution! Array parameters to function calls must be 1-based.
The basic syntax is the following, the <...> means optional parameters:
catXFi(format, delimiter, A)Arguments description:
-
format- A name of the numeric format to be used. -
delimiter- A delimiter string to be used. -
A- Numeric array
Example 1.
data _null_;
x = 1;
y = .;
z = 3;
array a[*] x y z;
length s $ 17;
s = catXFi("z5.", "#", A);
put (_all_) (=);
run;The catXFj() function is a wrapper of the catX() function but with ability to format character values.
For array A = ["a", " ", "c"] the
catXFj("upcase.", "*", A) returns "A**C"
If format does not handle nulls they are printed as an empty string.
Caution! Array parameters to function calls must be 1-based.
The basic syntax is the following, the <...> means optional parameters:
catXFj(format, delimiter, A)Arguments description:
-
format- A name of the character format to be used. -
delimiter- A delimiter string to be used. -
A- Character array
Example 1.
data _null_;
t = "t";
u = " ";
v = "v";
array b[*] t u v;
length s $ 17;
s = catXFj("upcase.", "*", B);
put (_all_) (=);
run;The catXFn() function is a wrapper
of the catX() function but with ability
to format numeric values.
For array A = [0, 1, 2] the
catXFn("date9.", "#", A) returns
"01JAN1960#02JAN1960#03JAN1960"
Caution! Array parameters to function calls must be 1-based.
The basic syntax is the following, the <...> means optional parameters:
catXFn(format, delimiter, A)Arguments description:
-
format- A name of the numeric format to be used. -
delimiter- A delimiter string to be used. -
A- Numeric array
Example 1.
data _null_;
x = 1;
y = .;
z = 3;
array a[*] x y z;
length s $ 17;
s = catXFn("z5.", "#", A);
put (_all_) (=);
run;The delDataset() function is a "wrapper"
for the Fdelete() function.
delDataset() function uses a text string with
a dataset name as an argument.
Function checks for *.sas7bdat, *.sas7bndx,
and *.sas7bvew files and delete them.
Return code of 0 means dataset was deleted.
For compound library files are deleted from ALL locations!
Note: Currently only the BASE SAS engine datasets/views are deleted.
Tested on Windows and Linux. Not tested on Z/OS.
The basic syntax is the following, the <...> means optional parameters:
delDataset(lbds_)Arguments description:
lbds_- Required, character argument containing name of the dataset/view to be deleted. The_last_special name is honored.
EXAMPLE 1.
data TEST1 TEST2(index=(x));
x = 17;
run;
data TEST3 / view=TEST3;
set test1;
run;
data _null_;
p = delDataset("WORK.TEST1");
put p=;
p = delDataset("TEST2");
put p=;
p = delDataset("WORK.TEST3");
put p=;
run;Example 2.
data TEST4;
x=42;
run;
data _null_;
p = delDataset("_LAST_");
put p=;
run;Example 3.
options dlcreatedir;
libname user "%sysfunc(pathname(work))/user";
data TEST5;
x=42;
run;
data _null_;
p = delDataset("test5");
put p=;
run;
libname user clear;Example 4.
data TEST6;
x=42;
run;
%put *%sysfunc(delDataset(test6))*;Example 5.
options dlcreatedir;
libname L1 "%sysfunc(pathname(work))/L)1";
libname L2 "%sysfunc(pathname(work))/L(2";
libname L3 "%sysfunc(pathname(work))/L'3";
data L1.TEST7 L2.TEST7 L3.TEST7;
x=42;
run;
libname L12 ("%sysfunc(pathname(work))/L(1" "%sysfunc(pathname(work))/L)2");
libname L1L2 (L2 L3);
%put *%sysfunc(delDataset(L12.test7))*;
%put *%sysfunc(delDataset(L1L2.test7))*;The semicolonC() function is internal function used by the semicolon format. Returns character value of length 32767.
The basic syntax is the following:
semicolonC(X)Arguments description:
X- Character value.
The semicolonN() function is internal function used by the semicolon format. Returns character value of length 33.
The basic syntax is the following:
semicolonN(X)Arguments description:
X- Numeric value.
The bpklength format and informats uses the klength() function
to count the number of letters in a word. For empty string returns 0.
Example 1. Informats and format:
data work.count_letters;
input x $ 32.;
n = input (x, bpklength.);
c = input (x, $bpklength.);
format x $bpklength.;
cards;
ż
żó
żół
żółć
a
ab
abc
abcd
空
空手
空手道
1
12
123
1234
12345
123456
1234567
12345678
123456789
1234567890
;
run;
proc print data=work.count_letters;
run;Example 2. Format:
data _null_;
length x $ 32767;
do i = 32767/3 to 0 by -1111, 10 to 0 by -1;
x=repeat("空",i);
put x $bpklength.;
end;
run; The bplength format and informats use the lengthn() function
to count the number of bytes in a word.
Example 1. Informats and format:
data work.count_bytes;
input x $ 32.;
n = input (x, bplength.);
c = input (x, $bplength.);
format x $bplength.;
cards;
ż
żó
żół
żółć
a
ab
abc
abcd
空
空手
空手道
1
12
123
1234
12345
123456
1234567
12345678
123456789
1234567890
;
run;
proc print data=work.count_bytes;
run;Example 2. Format:
data _null_;
length x $ 32767;
do i = 32767/3 to 0 by -1111, 10 to 0 by -1;
x=repeat("空",i);
put x $bplength.;
end;
run; The brackets format adds brackets around a text or a number. Leading and trailing spaces are dropped before adding brackets.
Example 1.
data _null_;
input x;
if x < 0 then put x= brackets.;
else put x= best32.;
cards;
2
1
0
-1
-2
;
run;The semicolon format adds semicolon after text or number. Leading and trailing spaces are dropped before adding semicolon.
Example 1.
data _null_;
x = 1;
y = "A";
put x= semicolon. y= $semicolon.;
run;The qsortInCbyProcProto() is external C function, this is the implementation of the Quick Sort algorithm.
The function is used internally by functions in the BasePlus package.
Asumptions:
- smaller subarray is sorted first,
- subarrays of size < 11 are sorted by insertion sort,
- pivot is selected as median of low index value, high index value, and (low+high)/2 index value.
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!CAUTION! Sorted array CANNOT contains SAS missing values !
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
The basic syntax is the following:
qsortInCbyProcProto(arr, low, high)Arguments description:
-
arr- An array of double type to be sorted. -
low- An integer low index of starting position (from which the sorting is done). -
high- An integer high index of ending position (up to which the sorting is done).
Reference 1.
Insertion sort for arrays smaller then 11 elements:
Based on the code from the following WikiBooks page [2020.08.14]:
https://pl.wikibooks.org/wiki/Kody_%C5%BAr%C3%B3d%C5%82owe/Sortowanie_przez_wstawianie
Reference 2.
Iterative Quick Sort:
Based on the code from the following pages [2020.08.14]:
https://www.geeksforgeeks.org/iterative-quick-sort/
https://www.geeksforgeeks.org/c-program-for-iterative-quick-sort/
The fromMissingToNumberBS() function gets numeric missing value or a number as an argument and returns an integer from 1 to 29.
For a numeric missing argument the returned values are:
- 1 for
._ - 2 for
. - 3 for
.a - ...
- 28 for
.zand - 29 for all other.
The function is used internally by functions in the BasePlus package.
For missing value arguments the function
is an inverse of the fromNumberToMissing() function.
The basic syntax is the following, the <...> means optional parameters:
fromMissingToNumberBS(x)Arguments description:
x- A numeric missing value or a number.
EXAMPLE 1.
data _null_;
do x = ._, ., .a, .b, .c, 42;
y = fromMissingToNumberBS(x);
put x= y=;
end;
run;The fromNumberToMissing() function gets a number as an argument and returns a numeric missing value or zero.
For a numeric argument the returned values are:
._for 1.for 2.afor 3- ...
.zfor 28 and0for all other.
The function is used internally by functions in the BasePlus package.
For arguments 1,2,3, ..., and 28 the function
is an inverse of the fromMissingToNumberBS() function.
The basic syntax is the following, the <...> means optional parameters:
fromNumberToMissing(x)Arguments description:
x- A numeric value.
EXAMPLE 1.
data _null_;
do x = 1 to 29;
y = fromNumberToMissing(x);
put x= y=;
end;
run;The quickSort4NotMiss() subroutine is an alternative to the
CALL SORTN() subroutine for 1-based big arrays (i.e. > 10'000'000 elements)
when memory used by call sortn() may be an issue.
For smaller arrays the memory footprint is not significant.
The subroutine is based on an iterative quick sort algorithm
implemented in the qsortInCbyProcProto() C prototype function.
Caution 1! Array CANNOT contains missing values!
Caution 2! Array parameters to subroutine calls must be 1-based.
The basic syntax is the following, the <...> means optional parameters:
call quickSort4NotMiss(A)Arguments description:
A- Argument is a 1-based array of NOT missing numeric values.
EXAMPLE 1. For session with 8GB of RAM, array of size 250'000'000 with values in range from 0 to 99'999'999 and NO missing values.
%let size = 250000000;
options fullstimer;
data _null_;
array test[&size.] _temporary_ ;
t = time();
call streaminit(123);
do _N_ = &size. to 1 by -1;
test[_N_] = int(100000000*rand("uniform"));
end;
t = time() - t;
put "Array population time: " t;
put "First 50 elements before sorting:";
do _N_ = 1 to 20;
put test[_N_] = @;
end;
t = time();
call quickSort4NotMiss (test);
t = time()-t;
put "Sorting time: " / t=;
put; put "First 50 elements after sorting:";
do _N_ = 1 to 20;
put test[_N_] = @;
end;
run;Example 2. Resources comparison for session with 8GB of RAM.
Array of size 250'000'000 with random values from 0 to 999'999'999 and NO missing values.
Array:
Population time 8.82s
memory 1'953'470.62k
OS Memory 1'977'436.00k
Call quickSort4NotMiss:
Sorting time 66.92s
Memory 1'954'683.06k
OS Memory 1'977'436.00k
Call quickSortLight:
Sorting time 70.98s
Memory 1'955'479.71k
OS Memory 1'977'436.00kThe quickSortHash() subroutine is an alternative to the
CALL SORTN() subroutine for 1-based big arrays (i.e. > 10'000'000 elements)
when memory used by call sortn() may be an issue.
For smaller arrays the memory footprint is not significant.
The subroutine is based on an iterative quick sort algorithm
implemented in the qsortInCbyProcProto() C prototype function.
The number of "sparse distinct data values" is set to 100'000 to
use the hash sort instead of the quick sort.
E.g. when number of unique values for sorting is less then
100'000 then an ordered hash table is used to store the data
and their count and sort them.
Caution! Array parameters to subroutine calls must be 1-based.
Note! Due to improper memory reporting/releasing for hash tables in FCMP procedure the reported memory used after running the function may not be in line with the RAM memory required for processing.
The basic syntax is the following, the <...> means optional parameters:
call quickSortHash(A)Arguments description:
A- Argument is a 1-based array of numeric values.
EXAMPLE 1. For session with 8GB of RAM Array of size 250'000'000 with values in range from 0 to 99'999'999 and around 10% of various missing values.
%let size = 250000000;
options fullstimer;
data _null_;
array test[&size.] _temporary_ ;
array m[0:27] _temporary_
(._ . .A .B .C .D .E .F .G .H .I .J .K .L
.M .N .O .P .Q .R .S .T .U .V .W .X .Y .Z);
t = time();
call streaminit(123);
do _N_ = &size. to 1 by -1;
_I_ + 1;
if rand("uniform") > 0.1 then test[_I_] = int(100000000*rand("uniform"));
else test[_I_] = m[mod(_N_,28)];
end;
t = time() - t;
put "Array population time: " t;
put "First 50 elements before sorting:";
do _N_ = 1 to 20;
put test[_N_] = @;
end;
t = time();
call quickSortHash (test);
t = time()-t;
put "Sorting time: " / t=;
put; put "First 50 elements after sorting:";
do _N_ = 1 to 20;
put test[_N_] = @;
end;
run;Example 2. For session with 8GB of RAM Array of size 250'000'000 with values in range from 0 to 9'999 and around 10% of various missing values.
%let size = 250000000;
options fullstimer;
data _null_;
array test[&size.] _temporary_ ;
array m[0:27] _temporary_
(._ . .A .B .C .D .E .F .G .H .I .J .K .L
.M .N .O .P .Q .R .S .T .U .V .W .X .Y .Z);
t = time();
call streaminit(123);
do _N_ = &size. to 1 by -1;
_I_ + 1;
if rand("uniform") > 0.1 then test[_I_] = int(10000*rand("uniform"));
else test[_I_] = m[mod(_N_,28)];
end;
t = time() - t;
put "Array population time: " t;
put "First 50 elements before sorting:";
do _N_ = 1 to 20;
put test[_N_] = @;
end;
t = time();
call quickSortHash (test);
t = time()-t;
put "Sorting time: " / t=;
put; put "First 50 elements after sorting:";
do _N_ = 1 to 20;
put test[_N_] = @;
end;
run;Example 3. Resources comparison for session with 8GB of RAM
A) Array of size 10'000'000 with random values from 0 to 9'999 range (sparse) and around 10% of missing data.
Array:
Population time 0.61s
Memory 78'468.50k
OS Memory 101'668.00k
Call sortn:
Sorting time 0.87s
Memory 1'120'261.53k
OS Memory 1'244'968.00k
Call quickSortHash:
Sorting time 6.76s
Memory 1'222'242.75k(*)
OS Memory 1'402'920.00k(*)
Call quickSortLight:
Sorting time 23.45s
Memory 80'527.75k
OS Memory 101'924.00kB) Array of size 10'000'000 with random values from 0 to 99'999'999 range (dense) and around 10% of missing data.
Array:
Population time 0.6s
Memory 78'463.65k
OS Memory 101'924.00k
Call sortn:
Sorting time 1.51s
Memory 1'120'253.53k
OS Memory 1'244'968.00k
Call quickSortHash:
Sorting time 6.28s
Memory 1'222'241.93k(*)
OS Memory 1'402'920.00k(*)
Call quickSortLight:
Sorting time 0.78s
Memory 80'669.28k
OS Memory 102'436.00kC) Array of size 250'000'000 with random values from 0 to 999'999'999 range (dense) and around 10% of missing data.
Array:
Population time 15.34s
memory 1'953'471.81k
OS Memory 1'977'436.00k
Call sortn:
FATAL: Insufficient memory to execute DATA step program.
Aborted during the COMPILATION phase.
ERROR: The SAS System stopped processing this step
because of insufficient memory.
Call quickSortHash:
Sorting time 124.68s
Memory 7'573'720.34k(*)
OS Memory 8'388'448.00k(*)
Call quickSortLight:
Sorting time 72.41s
Memory 1'955'520.78k
OS Memory 1'977'180.00kD) Array of size 250'000'000 with random values from 0 to 99'999 range (sparse) and around 10% of missing data.
Array:
Population time 16.07
Memory 1'953'469.78k
OS Memory 1'977'180.00k
Call sortn:
FATAL: Insufficient memory to execute DATA step program.
Aborted during the COMPILATION phase.
ERROR: The SAS System stopped processing this step
because of insufficient memory.
Call quickSortHash:
Sorting time 123.5s
Memory 7'573'722.03k
OS Memory 8'388'448.00k
Call quickSortLight:
Sorting time 1'338.25s
Memory 1'955'529.90k
OS Memory 1'977'436.00k(*) When using hash tables in Proc FCMP the RAM
usage is not indicated properly. The memory
allocation is reported up to the session limit
and then reused if needed. The really required
memory is in fact much less then reported.
The quickSortHashSDDV() subroutine is an alternative to the
CALL SORTN() subroutine for 1-based big arrays (i.e. > 10'000'000 elements)
when memory used by call sortn() may be an issue.
For smaller arrays the memory footprint is not significant.
The subroutine is based on an iterative quick sort algorithm
implemented in the qsortInCbyProcProto() C prototype function.
The number of "sparse distinct data values" (argument SDDV) may
be adjusted to use the hash sort instead of the quick sort.
E.g. when number of unique values for sorting is less then
some N then an ordered hash table is used to store the data
and their count and sort them.
Caution! Array parameters to subroutine calls must be 1-based.
Note! Due to improper memory reporting/releasing for hash tables in FCMP procedure the report memory used after running the function may not be in line with the RAM memory required for processing.
The basic syntax is the following, the <...> means optional parameters:
call quickSortHashSDDV(A, SDDV)Arguments description:
-
A- Argument is a 1-based array of numeric values. -
SDDV- A number of distinct data values, e.g. 100'000.
EXAMPLE 1. For session with 8GB of RAM Array of size 250'000'000 with values in range from 0 to 99'999'999 and around 10% of various missing values.
%let size = 250000000;
options fullstimer;
data _null_;
array test[&size.] _temporary_ ;
array m[0:27] _temporary_
(._ . .A .B .C .D .E .F .G .H .I .J .K .L
.M .N .O .P .Q .R .S .T .U .V .W .X .Y .Z);
t = time();
call streaminit(123);
do _N_ = &size. to 1 by -1;
_I_ + 1;
if rand("uniform") > 0.1 then test[_I_] = int(100000000*rand("uniform"));
else test[_I_] = m[mod(_N_,28)];
end;
t = time() - t;
put "Array population time: " t;
put "First 50 elements before sorting:";
do _N_ = 1 to 20;
put test[_N_] = @;
end;
t = time();
call quickSortHashSDDV (test, 2e4);
t = time()-t;
put "Sorting time: " / t=;
put; put "First 50 elements after sorting:";
do _N_ = 1 to 20;
put test[_N_] = @;
end;
run;Example 2. For session with 8GB of RAM Array of size 250'000'000 with values in range from 0 to 9'999 and around 10% of various missing values.
%let size = 250000000;
options fullstimer;
data _null_;
array test[&size.] _temporary_ ;
array m[0:27] _temporary_
(._ . .A .B .C .D .E .F .G .H .I .J .K .L
.M .N .O .P .Q .R .S .T .U .V .W .X .Y .Z);
t = time();
call streaminit(123);
do _N_ = &size. to 1 by -1;
_I_ + 1;
if rand("uniform") > 0.1 then test[_I_] = int(10000*rand("uniform"));
else test[_I_] = m[mod(_N_,28)];
end;
t = time() - t;
put "Array population time: " t;
put "First 50 elements before sorting:";
do _N_ = 1 to 20;
put test[_N_] = @;
end;
t = time();
call quickSortHashSDDV (test, 2e4);
t = time()-t;
put "Sorting time: " / t=;
put; put "First 50 elements after sorting:";
do _N_ = 1 to 20;
put test[_N_] = @;
end;
run;The quickSortLight() subroutine is an alternative to the
CALL SORTN() subroutine for 1-based big arrays (i.e. > 10'000'000 elements)
when memory used by call sortn() may be an issue.
For smaller arrays the memory footprint is not significant.
The subroutine is based on an iterative quick sort algorithm
implemented in the qsortInCbyProcProto() C prototype function.
Caution! Array parameters to subroutine calls must be 1-based.
The basic syntax is the following, the <...> means optional parameters:
call quickSortLight(A)Arguments description:
A- Argument is a 1-based array of numeric values.
EXAMPLE 1. For session with 8GB of RAM Array of size 250'000'000 with values in range from 0 to 99'999'999 and around 10% of various missing values.
%let size = 250000000;
options fullstimer;
data _null_;
array test[&size.] _temporary_ ;
array m[0:27] _temporary_
(._ . .A .B .C .D .E .F .G .H .I .J .K .L
.M .N .O .P .Q .R .S .T .U .V .W .X .Y .Z);
t = time();
call streaminit(123);
do _N_ = &size. to 1 by -1;
_I_ + 1;
if rand("uniform") > 0.1 then test[_I_] = int(100000000*rand("uniform"));
else test[_I_] = m[mod(_N_,28)];
end;
t = time() - t;
put "Array population time: " t;
put "First 50 elements before sorting:";
do _N_ = 1 to 20;
put test[_N_] = @;
end;
t = time();
call quickSortLight (test);
t = time()-t;
put "Sorting time: " / t=;
put; put "First 50 elements after sorting:";
do _N_ = 1 to 20;
put test[_N_] = @;
end;
run;Example 2. Resources comparison for session with 8GB of RAM.
Array of size 250'000'000 with random values from 0 to 999'999'999 and NO missing values.
Array:
Population time 8.82s
memory 1'953'470.62k
OS Memory 1'977'436.00k
Call quickSort4NotMiss:
Sorting time 66.92s
Memory 1'954'683.06k
OS Memory 1'977'436.00k
Call quickSortLight:
Sorting time 70.98s
Memory 1'955'479.71k
OS Memory 1'977'436.00kExample 3. Resources comparison for session with 8GB of RAM
A) Array of size 10'000'000 with random values from 0 to 9'999 range (sparse) and around 10% of missing data.
Array:
Population time 0.61s
Memory 78'468.50k
OS Memory 101'668.00k
Call sortn:
Sorting time 0.87s
Memory 1'120'261.53k
OS Memory 1'244'968.00k
Call quickSortHash:
Sorting time 6.76s
Memory 1'222'242.75k(*)
OS Memory 1'402'920.00k(*)
Call quickSortLight:
Sorting time 23.45s
Memory 80'527.75k
OS Memory 101'924.00kB) Array of size 10'000'000 with random values from 0 to 99'999'999 range (dense) and around 10% of missing data.
Array:
Population time 0.6s
Memory 78'463.65k
OS Memory 101'924.00k
Call sortn:
Sorting time 1.51s
Memory 1'120'253.53k
OS Memory 1'244'968.00k
Call quickSortHash:
Sorting time 6.28s
Memory 1'222'241.93k(*)
OS Memory 1'402'920.00k(*)
Call quickSortLight:
Sorting time 0.78s
Memory 80'669.28k
OS Memory 102'436.00kC) Array of size 250'000'000 with random values from 0 to 999'999'999 range (dense) and around 10% of missing data.
Array:
Population time 15.34s
memory 1'953'471.81k
OS Memory 1'977'436.00k
Call sortn:
FATAL: Insufficient memory to execute DATA step program.
Aborted during the COMPILATION phase.
ERROR: The SAS System stopped processing this step
because of insufficient memory.
Call quickSortHash:
Sorting time 124.68s
Memory 7'573'720.34k(*)
OS Memory 8'388'448.00k(*)
Call quickSortLight:
Sorting time 72.41s
Memory 1'955'520.78k
OS Memory 1'977'180.00kD) Array of size 250'000'000 with random values from 0 to 99'999 range (sparse) and around 10% of missing data.
Array:
Population time 16.07
Memory 1'953'469.78k
OS Memory 1'977'180.00k
Call sortn:
FATAL: Insufficient memory to execute DATA step program.
Aborted during the COMPILATION phase.
ERROR: The SAS System stopped processing this step
because of insufficient memory.
Call quickSortHash:
Sorting time 123.5s
Memory 7'573'722.03k
OS Memory 8'388'448.00k
Call quickSortLight:
Sorting time 1'338.25s
Memory 1'955'529.90k
OS Memory 1'977'436.00k(*) When using hash tables in Proc FCMP the RAM
usage is not indicated properly. The memory
allocation is reported up to the session limit
and then reused if needed. The really required
memory is in fact much less then reported.
The date() macro function is a "lazy typer" wrapping up %sysfunc(date()).
See examples below for the details.
The %date() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%date(<format>)Arguments description:
format- Optional, if a value is provided it should be a valid SAS format capable of handling values produced by thedate()function.
EXAMPLE 1. Get value of date():
%put %date();EXAMPLE 2. Get value of date() with a format:
%put %date(date11.);The datetime() macro function is a "lazy typer" wrapping up %sysfunc(datetime()).
See examples below for the details.
The %datetime() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%datetime(<format>)Arguments description:
format- Optional, if a value is provided it should be a valid SAS format capable of handling values produced by thedatetime()function.
EXAMPLE 1. Get value of datetime():
%put %datetime();EXAMPLE 1. Get value of datetime() as "long" and "short" ISO-8601:
%put %datetime(e8601dt.);
%put %datetime(b8601dt.);EXAMPLE 1. Get value of datetime() with user defined format:
proc format;
picture myCrazyDT (default=50)
other='%0Ssec. %0Mmin. %0Hhour %0dday %0mmonth %Yyear' (datatype=datetime)
;
run;
%put %datetime(myCrazyDT.);The downloadFilesTo() macro copy files (in binary mode
using filename() function with options lrecl=1 recfm=n)
from list provided by user to a directory indicated
in the macro call.
Macro can be executed in two possible ways:
- by providing list of files to download in a
datalines4(cards4) list directly after macro call:
%downloadFilesTo(</path/to/target/directory>)
datalines4;
<link to file1>
<link to file2>
...
<link to fileN>
;;;;
run;- by create a dataset with a list of links and use of
DS=,DSvar=, andDSout=parameters:
%downloadFilesTo(</path/to/target/directory>
, DS=<dataset with list>
, DSvar=<variable with list>
, DSout=<variable with names for downloaded files>
)See examples below for the details.
The %downloadFilesTo() macro does not execute as a pure macro code.
Temporary dataset work.______locationInfoData is generated during processing.
The basic syntax is the following, the <...> means optional parameters:
%downloadFilesTo(
target
<,DS=>
<,DSvar=link>
<,DSout=scan(link,-1,"/\")>
<,inDev=URL>
<,outDev=DISK>
<,inOptions=>
<,outOptions=>
)Arguments description:
target- Required, a path to target directory. If empty theWORKlocation is used.
*. DS= - Optional, name of data set with list
of files to download.
*. DSvar= - Optional, name of variable in data set
with list of files to download.
*. DSout= - Optional, name of variable in data set
with list of names for to downloaded files.
Default value is: scan(link,-1,"/\") it is
an expression to cut last part of the link.
*. inDev= - Optional, type of device used by the
filename() function to access incoming files.
Default value is URL.
*. outDev= - Optional, type of device used by the
filename() function to access outgoing files.
Default value is DISK.
*. inOptions= - Optional, list of additional options for the
filename() function to access incoming files.
Default value is empty.
*. outOptions= - Optional, list of additional options for the
filename() function to access outgoing files.
Default value is empty.
EXAMPLE 1. Download data from web with direct list and then copy between directories:
resetline;
%downloadFilesTo(~/directoryA)
datalines4;
https://www.lexjansen.com/wuss/2023/WUSS-2023-Paper-189.pdf
https://www.lexjansen.com/wuss/2023/WUSS-2023-Paper-189.zip
;;;;
run;
%downloadFilesTo(~/directoryB,inDev=DISK)
datalines4;
~/directoryA/WUSS-2023-Paper-189.pdf
~/directoryA/WUSS-2023-Paper-189.zip
;;;;
run;EXAMPLE 2. Download data from web using data set with list:
resetline;
data listOfFiles;
infile cards;
input files :$1024.;
cards4;
https://www.lexjansen.com/wuss/2023/WUSS-2023-Paper-201.pdf
https://www.lexjansen.com/wuss/2023/WUSS-2023-Paper-109.pdf
;;;;
run;
%downloadFilesTo(R:\directoryC, DS=listOfFiles, DSvar=files)The %expandDataSetsList() macro is dedicated to "lazy typers".
It allows to expand data set names provides in form of a SAS list for a given library.
See examples below for the details.
The %expandDataSetsList() macro works as pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%expandDataSetsList(
lib
,datasets
<,quote=>
<,views=>
)Arguments description:
-
lib- Required, is a name of a library where data sets are looked-up. -
datasets- Required, is a list of data sets or views to be expanded. Can be named list (e.g.x_:), can be enumerated list (e.g.y_1-y_5), or both. Also the_ALL_value is accepted.
*. quote - Optional, binary indicator (default 0 means "no").
Tells if data set names should be quoted.
*. views - Optional, binary indicator (default 1 means "yes").
Tells if views names should be listed too.
EXAMPLE 0. Create data sets for tests:
data x_a x_b x_c x_d x_e x_f;
x=17;
run;
data y1 y2 y3 y4 y9 y10 y11 y12;
y=42;
run;EXAMPLE 1. Display names of listed datasets, handle not existing too:
%PUT #%expandDataSetsList(lib=work,datasets=x_: y1-y4 y9)#;
%PUT #%expandDataSetsList(lib=work,datasets=x_: y1-y4 notExist_1-notExist_10)#;EXAMPLE 2. Display one and all datasets from WORK, names quoted:
%PUT %expandDataSetsList(lib=work,datasets=_ALL_, quote=1);
%PUT %expandDataSetsList(lib=work,datasets=x_a, quote=1);EXAMPLE 3. In tandem with the resolve() function:
data _null_;
test=resolve('%expandDataSetsList(lib=work,datasets=x_:, quote=1)');
put test / test hex72.;
run;EXAMPLE 4. Workaround for proc delete not working with colon operator,
observe a warning info for x_: list:
proc delete lib=work data=%expandDataSetsList(lib=work,datasets=x_:)
;
run;
%PUT #%expandDataSetsList(lib=work,datasets=x_: y1-y4 y9)#;EXAMPLE 5. Including and excluding views:
%PUT %expandDataSetsList(lib=sashelp,datasets=_ALL_);
%PUT %expandDataSetsList(lib=sashelp,datasets=_ALL_,views=0);The filePath() macro function returns path to a file,
it is a wrapper to pathname() function for files.
See examples below for the details.
The %filePath() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%filePath(
fileref
)Arguments description:
fileref- Required, a fileref from thefilenamestatement.
EXAMPLE 1. Return path to temporary file:
filename f temp;
%put %filePath(f);The findDSwithVarVal() macro searches for all datasets (available for a given session) containing a variable of a given value.
The value search is case sensitive - but can be altered with IC= parameter.
The value search keeps leading blanks - but can be altered with TB= parameter.
The value search compares full value - but can be altered with CTS= parameter.
The default variable type is char, the type= parameter allows
to change it (possible values are char and num), the parameter is case sensitive.
Only datasets are searched, views are not included.
During the process two temporary datasets named:
WORK._ (single underscore) and WORK.__ (double underscore)
are generated. The datasets are deleted at the end of the process.
By default search results are stored in the WORK.RESULT dataset.
Name of the dataset can be altered with result= parameter.
The dataset with result contains two variables:
datasetName - names of datasets,
firstObservation - the firs occurrence of the value.
See examples below for the details.
The %findDSwithVarVal() macro does not execute as a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%findDSwithVarVal(
variable
,value
<,type=>
<,ic=>
<,tb=>
<,cts=>
<,lib=>
<,result=>
)Arguments description:
-
variable- Required, name of variable to be searched. -
value- Required, the value to be searched.
*. type - Optional, default value is char.
Indicates which type is the searched value.
Possible values are char and num,
the parameter is case sensitive.
*. ic - Optional, "Ignore Cases", default value is 0.
Indicates should the search ignore cases of the text values.
Possible values are 0 and 1.
*. tb - Optional, "Trim Blanks", default value is 0.
Indicates should the search trim leading and trailing
blanks of the text values.
Possible values are 0 and 1.
*. cts - Optional, "Compare To Shorter", default value is 0.
IF set to 1 execute value comparison as =: for the text value.
Possible values are 0 and 1.
See examples.
*. lib - Optional, default value is missing.
If not empty narrows the search to a particular library.
*. result - Optional, default value is WORK.RESULT.
Is the name of the dataset with results.
EXAMPLE 1. Search variable NAME containing value John:
%findDSwithVarVal(name, John)EXAMPLE 2. Search numeric variable AGE containing value 14:
%findDSwithVarVal(age, 14, type=num)EXAMPLE 3. Search numeric variable SCORE with missing value:
data TEST;
score=17; output;
score=42; output;
score=. ; output;
run;
%findDSwithVarVal(score, ., type=num, result=WORK.MissingScore)EXAMPLE 4. Search library WORK for variable NAME starting with value Jo
ignoring cases and trimming blanks from value:
data A;
name="Joanna";
data B;
name="john";
data C;
name=" Joseph";
data D;
name=" joe";
run;
%findDSwithVarVal(name, Jo, ic=1, tb=1, cts=1, lib=WORK)The fmt() macro function returns a value formatted by a format,
it is a wrapper to putN() and putC() functions.
See examples below for the details.
The %fmt() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%fmt(
value
,format
,align
<,type=>
)Arguments description:
-
value- Required, a value to be formatted. -
format- Required, a name of a format to be used, character format should be preceded by the$. -
align- Optional, allows to use the-L,-Rand-Cmodifiers.
type=n- Optional, defines type of the format. If the format name is preceded by the$then C is set automatically. If the character format name is without$then set value toCyourself.
EXAMPLE 1. Formatting values:
%put %fmt(111, 7.2);
%put %fmt(111, dollar10.2);
%put %fmt(abc, $upcase.);
%put %fmt(12345, date9.);EXAMPLE 2. Align values (compare different results!):
%put *%fmt(ABC, $char9., -L)*;
%put *%fmt(ABC, $char9., -R)*;
%put *%fmt(ABC, $char9., -C)*;
%put %fmt(ABC, $char9., -L);
%put %fmt(ABC, $char9., -R);
%put %fmt(ABC, $char9., -C);The %GenerateOneLiners() macro is a "macro-generator" dedicated
to "lazy typers".
It allows to generate macro wrappers for functions that have the following form:
%macro FUNCTION()/parmbuff;
%sysfunc(FUNCTION&syspbuff)
%mend FUNCTION;
%macro qFUNCTION()/parmbuff;
%qsysfunc(FUNCTION&syspbuff)
%mend qFUNCTION;See examples below for the details.
The %GenerateOneLiners() macro is not pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%GenerateOneLiners(
<,listOfFunctions=>
<,prefix=>
)Arguments description:
-
listOfFunctions- Required, is a space separated list of valid SAS functions. Default value is:CATX CATQ CATT CAT COMPRESS REVERSE REPEAT. -
prefix- Optional, a prefix added to the name of a created macro.
EXAMPLE 1. Create list of macrofunctions for
CATX CATQ CATT CAT COMPRESS REVERSE REPEAT:
%GenerateOneLiners(prefix=_)
%let a = 1,2,3,4,5,6;
%put %_CATX(%str( ),&a.);
%put %_CATQ(2A,&a.);
%put %_QCATQ(1AMD,%str(,),&a.);
%let x=a 1 b 2 c 3 d 4 e 5 f 6 g;
%put %_COMPRESS(&x.);
%put %_COMPRESS(&x.,,ka);EXAMPLE 2. Create list of macrofunctions for
SUM MEAN MEDIAN VAR STD USS CSS RANGE IQR MAD SUMABS:
%GenerateOneLiners(
listOfFunctions=SUM MEAN MEDIAN VAR STD USS CSS RANGE IQR MAD SUMABS
, prefix=_)
%put
%_SUM(1,2,3,4,5,6)
%_MEAN(1,2,3,4,5,6)
%_MEDIAN(1,2,3,4,5,6)
%_VAR(1,2,3,4,5,6)
;EXAMPLE 3. Some other lists:
%GenerateOneLiners(prefix=_
, listOfFunctions=CDF PDF RAND QUANTILE SQUANTILE SDF logCDF logPDF logSDF RMS
)
%GenerateOneLiners(prefix=_
, listOfFunctions=YEAR QTR MONTH WEEK DAY HOUR MINUTE SECOND
)
%GenerateOneLiners(prefix=_
, listOfFunctions=PCTL1 PCTL2 PCTL3 PCTL4 PCTL5 PCTL
)
%GenerateOneLiners(prefix=_
, listOfFunctions=YYQ MDY HMS INTCK INTNX SLEEP
)
%GenerateOneLiners(prefix=_
, listOfFunctions=WHICHC WHICHN
)
%GenerateOneLiners(prefix=_
, listOfFunctions=SYMEXIST SYMGLOBL SYMLOCAL
)
%GenerateOneLiners(prefix=_
, listOfFunctions=PRXCHANGE PRXMATCH PRXPAREN PRXPARSE
)
%GenerateOneLiners(prefix=_
, listOfFunctions=MD5 SHA256 HASHING
)The getTitle() macro extract text of titles or footnotes into a delimited list.
Titles/footnotes numbers can be selected with the number argument.
Only the text of a title or footnote is extracted.
Author of the original code is: Quentin McMullen ([email protected]).
See examples below for the details.
The %getTitle() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%getTitle(
< number>
<,type=>
<,dlm=>
<,qt=>
)Arguments description:
number- Optional, default value is empty, indicates numbers of titles to be extracted. Space-separated list is expected. If empty or_ALL_extract all non-missing.
*. type - Optional, default value is T.
Indicates which type is the searched.
T fro title, F for footnote.
*. dlm - Optional, "DeLiMiter", default value is | (pipe).
Possible values are: | \ / , . ~ * # @ ! + - _ : ?
or s for space, c for comma, d for semicolon.
*. qt - Optional, "QuoTes", default value is empty.
Use %str() for single quote symbol (e.g. %str(%")).
If there are multiple symbols, only the first and the
second are selected as a leading and trailing one,
e.g. qt="" gives "title1 text" "title2 text" ... .
EXAMPLE 1. Get titles in different forms:
title1 j=c "Hi Roger" ;
title2 j=l "Good Morning" ;
title3 "How are you?" ;
title4 ;
title5 "Bye bye!" ;
%put %GetTitle() ;
%put %GetTitle(1 3,dlm=c, qt=[]) ;
%put %GetTitle(2:4,dlm=s, qt='') ;EXAMPLE 2. Get footnotes in different forms:
footnote1 "First little footnote";
footnote2 "Second little footnote";
footnote3 "Third little footnote";
%put %GetTitle(1 2,type=f,dlm=s, qt="") ;
%put %GetTitle(2 3,type=f,dlm=c, qt='') ;
%put %GetTitle(1 3,type=f,dlm=d, qt=[]) ;The iffunc() macro function is a "lazy typer" wrapping up conditional execution.
Instead typing:
%let x = 1;
%if (1=&x.) %then
%do;
%let test = TRUE;
%end;
%else
%do;
%let test = FALSE;
%end;
%put &=test.;you can make it:
%let x = 1;
%let test = %iffunc((1=&x.),TRUE,FALSE);
%put &=test.;The %iffunc() macro ca be nested in open code
in contrary to standard %IF-%THEN-%ELSE.
The minoperator option is used to handle IN operator,
but the IN operator in the condition has to be surrounded
by spaces, e.g. &x. in (1 2 3).
Condition of the form &x. in(1 2 3) (no space after in) will
result with an error.
See examples below for the details.
The %iffunc() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%iffunc(
cond
,true
,false
<,v=>
)Arguments description:
-
cond- Required, is a condition to be evaluated. It has to be provided in brackets, e.g.,(1=&x.)or(%scan(&x.,1)=A). -
true- Required, value returned when condition is true. -
false- Required, value returned when condition is false.
EXAMPLE 1. Simple test:
%let x = 1;
%let test = %iffunc((&x.=1),True,False);
%put &=test.;EXAMPLE 2. Verbose with v= in the %put statement:
%put %iffunc((TRUE=true),TRUE,FALSE,v=1);EXAMPLE 3. Condition can be with decimals or text:
%put True: %iffunc((9.9<10),TRUE,FALSE);
%put False: %iffunc((A>B),TRUE,FALSE);
%put True: %iffunc((1=1.0),TRUE,FALSE);
%put False: %iffunc((1/3 = 0.5),TRUE,FALSE);
%put True: %iffunc((.1+.1+.1+.1+.1+.1+.1+.1+.1+.1 = 1),TRUE,FALSE);EXAMPLE 4. Macro-Functions works too:
%let x = A B C;
%put %iffunc((%scan(&x.,1)=A),Starts with "A"., Does not start with "A".);EXAMPLE 5. Simple nesting (in open code):
%let x=0;
%let y=0;
%put
%iffunc((&x.)
,It is X.
,%iffunc((&y.)
,It is Y.
,It is Z.
,v=1)
,v=1)
;EXAMPLE 6. Special characters have to be masked:
%put %iffunc(1,%str(,),%str(;));
%put %iffunc(0,%str(,),%str(;));EXAMPLE 7. If you want to call macros there is a price.
The %nrstr() and %unquote() are required:
%macro doX(a,b);
%put >>> &a. &b;
data do_X;
set sashelp.class;
run;
%mend;
%macro doY(a,b);
%put ### &a. &b;
data do_Y;
set sashelp.cars;
run;
%mend;
%macro doZ(a,b);
%put @@@ &a. &b;
data do_Z;
set sashelp.iris;
run;
%mend;
%let x=0;
%let y=0;
%unquote(
%iffunc((&x.)
,%nrstr(%doX(1,2))
,%iffunc((&y.)
,%nrstr(%doY(3,4))
,%nrstr(%doZ(5,6))
,v=1)
,v=1)
)EXAMPLE 8. Traffic lights with NOTE: and ERROR:,
Remember to handle the IN operator with an extra care.
Spaces are required around the IN operator.
%let x = A;
%put %iffunc((&x. IN (A B C)),NOTE:,ERROR:) x=%superq(x);
%let x = 7;
%put %iffunc((&x. IN (1 3 5)),NOTE:,ERROR:) x=%superq(x);
%let x = 1.1;
%put %iffunc((&x. IN (1.1 2.1 3.1)),NOTE:,ERROR:) x=%superq(x);EXAMPLE 9. Mixing IN operator with non-integer evaluation
requires %sysevalf() function:
%let x = Y;
%let y = 9.9;
%put THIS will be False: %iffunc((&x. in (y Y) AND &y.<10),TRUE,FALSE);
%put THIS will be True: %iffunc((&x. in (y Y) AND %sysevalf(&y.<10)),TRUE,FALSE);The infmt() macro function returns a value read in by an informat,
it is a wrapper to inputN() and inputC() functions.
See examples below for the details.
The %infmt() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%infmt(
value
,informat
<,type=>
)Arguments description:
-
value- Required, a value to be formatted. -
informat- Required, a name of a format to be used, character format should be preceded by the$.
type=n- Optional, defines type of the informat. If the informat name is preceded by the$then C is set automatically. If the character format name is without$then set value toCyourself.
EXAMPLE 1. Informatting values:
%put %infmt(111, 7.2);
%put %infmt(111.234, 7.2);
%put %infmt($111, dollar10.2);
%put %infmt($111.234, dollar10.2);
%put %infmt(abc, $upcase.);
%put %infmt(12mar45, date9.);The letters() macro function allows to print a list of Roman
letters starting from start up to end incremented by by.
The letters list can be uppercases or lowercase (parameter c=U or c=L),
can be quoted (e.g. q="" or q=[]), and can be separated by s=.
Values of start, end, and by have to be integers in range between 1 ad 26.
See examples below for the details.
The %letters() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%letters(
range
<,c=>
<,q=>
<,s=>
)Arguments description:
range- Required, letters selector in formstart:end:by. Lists letters fromstarttoendbyby. Values ofstart,end, andbyare separated by colon and must be between 1 ad 26. If value is outside range it is set tostart=1,en=26, andby=1. Ifendis missing then is set to value ofstart. Ifendis smaller thanstartlist is reversed
-
c = U- Optional, it is a lowercase letters indicator. SelectLorl. Default value isUfor upcase. -
q =- Optional, it is a quite around elements of the list. Default value is empty. Use%str()for one quote symbol. If there are multiple symbols, only the first and the second are selected as a preceding and trailing one, e.g.q=[]gives[A] [B] ... [Z]. -
s = %str( )- Optional, it is a separator between elements of the list. Default value is space.
EXAMPLE 1. Space-separated list of capital letters from A to Z:
%put %letters(1:26:1);
%put %letters();EXAMPLE 2. First, thirteenth, and last letter:
%put %letters(1) %letters(13) %letters(26);EXAMPLE 3. Every third lowercase letter, i.e. a d g j m p s v y:
%put %letters(1:26:3,c=L);EXAMPLE 4. Lists with separators:
%put %letters(1:26:2,s=#);
%put %letters(1:26:3,s=%str(;));
%put %letters(1:26:4,s=%str(,));
%put %letters(1:26,s=);
%put %letters(1:26,s==);
%put %letters(1:26,s=/);EXAMPLE 5. Every second letter with quotes:
%put %letters(1:26:2,q=%str(%'));
%put %letters(2:26:2,q=%str(%"));
%put %letters(1:26:2,q='');
%put %letters(2:26:2,q="");
%put %letters(1:26:2,q=<>);
%put %letters(2:26:2,q=\/);
%put %letters(1:26:2,q=());
%put %letters(2:26:2,q=][);EXAMPLE 6. Mix of examples 4, 5, and 6:
%put %letters(1:26,c=L,q='',s=%str(, ));EXAMPLE 7. If end is smaller than start list is reversed:
%put %letters(26:1:2,q='');The libPath() macro function returns path to a library,
it is a wrapper to pathname() function for libraries.
See examples below for the details.
The %libPath() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%libPath(
libref
)Arguments description:
libref- Required, a libref from thelibnamestatement.
EXAMPLE 1. Return path to WORK library:
%put %libPath(WORK);EXAMPLE 2. Return path to SASHELP library:
%put %libPath(SASHELP);The mInclude() macro is a macrolanguage version of the SAS %include statement.
But it allows for "embedding any code anywhere into SAS programs".
Macro was inspired by Leonid Batkhan and his blog post:
"Embedding any code anywhere into SAS programs" from May 30, 2023.
Link: https://blogs.sas.com/content/sgf/2023/05/30/embedding-any-code-anywhere-into-sas-programs/
The implementation presented, in contrary to inspiration source, is
based on the doSubL() function and a list of global
macro variables of the form ______<N> (six underscores and a number).
See examples below for the details.
The %mInclude() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%mInclude(
< f>
<,source=>
<,lrecl=>
<,symdel=>
)Arguments description:
f- Required, a SASfilerefor a quoted path to the included file.
*. source=0 - Optional, default value is 0.
Set to 1 if the source should be printed in the log.
*. lrecl=32767 - Optional, default value is 32767.
Sets the lrecl value for the file width.
*. symdel=1 - Optional, default value is 1.
Indicates if the global macro variables
______1 to ______N should be deleted
when the macro ends.
EXAMPLE 1. Embedding text in statements (the %include won't work here):
resetline;
filename f "%workpath()/testFile1.txt";
filename f list;
data _null_;
file f;
put "13 14 15";
run;
resetline;
data testDataset;
set sashelp.class;
where age in ( %mInclude(f) );
run;
data testDataset2;
set sashelp.class;
where age in ( %mInclude(f,source=1) );
run;
filename f clear;EXAMPLE 2. Embedding with direct path (mind those quotes!):
resetline;
filename f "%workpath()/testFile2.txt";
filename f list;
%let someGlobalMacroVariable=17;
data _null_;
file f;
put "options mprint;";
do i=1 to 3;
put "data y; x = " i "; run;";
put '%macro A' i +(-1) '(); %put ' i ' ** &someGlobalMacroVariable.; %mend; %A' i +(-1) '()';
end;
put "options nomprint;";
run;
resetline;
%mInclude("%workpath()/testFile2.txt")
%mInclude("%workpath()/testFile2.txt",source=1)EXAMPLE 3. Embedding SQL code inside the pass through execution:
resetline;
filename f2 "%workpath()/testSql.txt";
data _null_;
file f2;
input;
put _infile_;
cards4;
select
c2.make
, c2.model
, c2.type
, c2.invoice
, c2.date
from
public.CARS_EU c2
where
c2.cylinders > 4
and
c2.date > '2023-04-02'
;;;;
run;
title 'the %include fails';
proc sql;
connect to POSTGRES as PSGDB (
server="127.0.0.1"
port=5432
user="user"
password="password"
database="DB"
);
select * from connection to PSGDB
(
%Include f2 / source2;
)
;
disconnect from PSGDB;
quit;
title 'the %mInclude works';
proc sql;
connect to POSTGRES as PSGDB (
server="127.0.0.1"
port=5432
user="user"
password="password"
database="DB"
);
select * from connection to PSGDB
(
%mInclude(f2, source=1)
)
;
disconnect from PSGDB;
quit;
title;EXAMPLE 4. In a limited way and with help of the resolve() function,
it even works with IML's interface to R:
resetline;
filename f3 TEMP;
data _null_;
file f3;
infile cards4;
input;
put _infile_ ';'; %* a "semicolon" trick for R statements separation *;
cards4;
rModel <- lm(Weight ~ Height, data=Class, na.action="na.exclude")
print (rModel$call)
print (rModel)
;;;;
run;
proc iml;
codeText = resolve(' %mInclude(f3, source=1) ');
print codeText;
call ExportDataSetToR("Sashelp.Class", "Class" );
submit codeText / R;
&codeText
endsubmit;
quit;
The monthShift() macro is a utility macro which allows to shift "year-month" period by a given number of "periods" (months).
The result is in the YYYYMM format but can be altered.
See examples below for the details.
The %monthShift() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%monthShift(
< Y>
<,M>
<,shift>
<,ofmt=>
)Arguments description:
-
Y- Optional, a year from which counting starts. If null the value is set to system year. -
M- Optional, a month from which counting starts. If null the value is set to1. Can be a number (1to12) or a name (June,OCTOBER) or a three letters short (JAN,apr). -
shift- Optional, number of periods to shift. If null the value is set to0. Positive value shifts to the "future", negative value shifts to the "past", Can be an expression (e.g.1+2*3, see examples).
ofmt=YYMMn6.- Optional, it is a format name used to display the result. Default value isYYMMn6.See examples.
EXAMPLE 1. Shift one up and one down:
%put
Past: %monthShift(2023, 1, -1)
Current: %monthShift(2023, 1 )
Future: %monthShift(2023, 1, +1)
;EXAMPLE 2. Shift by expression:
%let n = 2;
%put
%monthShift(2023, 1, +1 + &n.*3)
;EXAMPLE 3. Shift with default values:
%put %monthShift();
%put %monthShift(2023);
%put %monthShift(2023,Jan);EXAMPLE 4. Shift with months names:
%put
%monthShift(2023,Jan,-1)
%monthShift(2023,Feb,-2)
%monthShift(2023,Mar,-3)
%monthShift(2023,Apr,-4)
%monthShift(2023,May,-5)
%monthShift(2023,Jun,-6)
%monthShift(2023,Jul,-7)
%monthShift(2023,Aug,-8)
%monthShift(2023,Sep,-9)
%monthShift(2023,Oct,-10)
%monthShift(2023,Nov,-11)
%monthShift(2023,Dec,-12)
;
%put
%monthShift(2023,January,12)
%monthShift(2023,February,11)
%monthShift(2023,March,10)
%monthShift(2023,April,9)
%monthShift(2023,May,8)
%monthShift(2023,June,7)
%monthShift(2023,July,6)
%monthShift(2023,August,5)
%monthShift(2023,September,4)
%monthShift(2023,October,3)
%monthShift(2023,November,2)
%monthShift(2023,December,1)
;EXAMPLE 5. Play with formatting:
%put
%monthShift(2023, 1, +1 )
%monthShift(2023, 1, +1, ofmt=yymm7. )
%monthShift(2023, 1, +1, ofmt=yymmd7.)
%monthShift(2023, 1, +1, ofmt=yymms7.)
;EXAMPLE 6. Read monthly data with noDSNFERR option:
data
A202210 A202211 A202212
A202301 A202302 A202303
A202304 A202305 A202306
A202307 A202308 A202309
;
set sashelp.class;
run;
options noDSNFERR;
data ALL;
set
A%monthShift(2023, 9, -12) - A%monthShift(2023, 9)
;
run;
options DSNFERR;The repList() macro function allows to repeat T
times elements of a L list, possibly E times each element,
separated by string S.
See examples below for the details.
The %repList() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%repList(
list
<,times=>
<,each=>
<,lenghtOut=>
<,sep=>
)Arguments description:
list- Required, a list of elements to be repeated. List can be space or comma-separated. Elements can be in quotes. For comma-separated list add brackets e.g.,%repList((A,B,C,D),times=5). The list separators are:<{[( ,;)]}>.
-
times=- Optional, An integer indicating the number of repetitions. By default set to1. -
each=- Optional, A list of integers indicating the number of repetitions of each element of the list e.g., for a listA B Cand theeach=2 4the result isA A B B B B C C. If the number of integers is less then the length of the list values are recycled from the beginning. By default set to1. -
lenghtOut=- Optional, An integer indicating after what the number of repetitions process will stop. By default set to0which means "do not stop". -
sep=- Optional, it is a separator printed between repeated elements. Mnemonics for space iss, for comma isc, and for semicolon inq. Default value is a single space.
EXAMPLE 1. Simple repetition of all elements:
%put %repList((A,B,C,D), times=3);EXAMPLE 2. Simple repetition of each element:
%put %repList(("A",'B',"C",'D'), each=3);EXAMPLE 3. Simple repetition with a separator:
%put %repList(A10;B20;C30, times=3, each=2, sep=Q);EXAMPLE 4. Recycle elements up to 8 with a comma as a separator:
%put %repList(1 2 3, lenghtOut=8, sep=c);EXAMPLE 5. Separate number of repetitions for each element:
%put [%repList([D][C][B][A], each = 2 3 5 7, sep=] [)];EXAMPLE 6. "ASCII art" butterflies:
%put {>%repList(! $ |, times = 2, each =2 1, sep=<} ... {>)<};EXAMPLE 7. Data repeating:
data A;
x=17;
data B;
x=42;
data C;
x=303;
run;
data Times2_A10B11C12;
set
%repList(A B C, times = 2, each =10 11 12)
;
run;The time() macro function is a "lazy typer" wrapping up %sysfunc(time()).
See examples below for the details.
The %time() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%time(<format>)Arguments description:
format- Optional, if a value is provided it should be a valid SAS format capable of handling values produced by thetime()function.
EXAMPLE 1. Get value of time():
%put %time();EXAMPLE 2. Get value of time() with a format:
%put %time(time8.);The today() macro function is a "lazy typer" wrapping up %sysfunc(today()).
See examples below for the details.
The %today() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%today(<format>)Arguments description:
format- Optional, if a value is provided it should be a valid SAS format capable of handling values produced by thetoday()function.
EXAMPLE 1. Get value of today():
%put %today();EXAMPLE 2. Get value of today() with a format:
%put %today(yymmdd10.);The translate() macro function allows to replace bytes with bytes in text string.
See examples below for the details.
The %translate() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%translate(
string
,from
,to
)Arguments description:
-
string- Required, string to modify. -
from- Required, list of bytes to be replaced with corresponding bytes fromto. -
to- Required, list of bytes replacing corresponding bytes fromfrom.
EXAMPLE 1. Replace quotes and commas with apostrophes and spaces:
%put %translate(%str("A", "B", "C"),%str(%",),%str(%' ));EXAMPLE 2. Unify all brackets;
%put %translate(%str([A] {B} (C) <D>),{[(<>)]},(((()))));EXAMPLE 3. Replace all digits with *:
%put %translate(QAZ1WSSX2EDC3RFV4TGB5YHN6UJM7IK8OL9P0,1234567890,**********);EXAMPLE 4. Letters change:
%put %translate(%str(A=B),AB,BA);The tranwrd() macro function allows to replace substrings with other substrings in text string.
Returned string is unquoted by %unquote().
See examples below for the details.
The %tranwrd() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%tranwrd(
string
,from
,to
<,repeat>
)Arguments description:
-
string- Required, string to modify. -
from- Required, substring replaced with corresponding string fromto. -
to- Required, substring replacing corresponding substring fromfrom. -
repeat- Optional, number of times the replacing should be repeated, default is 1. Useful while removing multiple adjacent characters, e.g. compress all multiple spaces (see example 2).
EXAMPLE 1. Simple text replacement:
%put %tranwrd(Miss Joan Smith,Miss,Ms.);EXAMPLE 2. Delete multiple spaces;
%put %tranwrd(%str(A B C),%str( ),%str( ),5);EXAMPLE 3. Remove substring:
%put %tranwrd(ABCxyzABCABCxyzABC,ABC);The %unifyVarsCaseSize() macro converts all variables names into low-case or
upcase letters for given library and list of datasets. Only necessary conversion is
done, i.e., variable abc will not be converted to low-case letters.
See examples below for the details.
The %unifyVarsCaseSize() macro works as pure macro code.
[NOTE:] The macro internally uses the %expandDataSetsList() macro.
The basic syntax is the following, the <...> means optional parameters:
%unifyVarsCaseSize(
lib
,ds
<,case=>
<,debug=>
)Arguments description:
-
lib- Required, is a name of a library where data sets are looked-up. -
ds- Required, is a list of data sets to be expanded. Can be named list (e.g.x_:), can be enumerated list (e.g.y_1-y_5), or both. Also the_ALL_value is accepted.
*. case - Optional, single letter indicator (default L means "low-case").
Tells if variables names should low-cased (l,L) or upcased ("u", "U").
*. debug - Optional, binary indicator (default 0 means "no").
Tells if processing notes should be printed.
EXAMPLE 0. Create data sets for tests:
data aclass1 aclass2 aclass3 bclass4 bclass5 bclass6;
set sashelp.class(obs=6);
Nn=_N_;
if 1=_N_ then output aclass1;
if 2=_N_ then output aclass2;
if 3=_N_ then output aclass3;
if 4=_N_ then output bclass4;
if 5=_N_ then output bclass5;
if 6=_N_ then output bclass6;
run;
proc print data=aclass1;
run;
proc print data=bclass6;
run;EXAMPLE 1. Convert all variables names to low-case:
%unifyVarsCaseSize(work,aclass:)
proc print data=aclass1;
proc print data=aclass2;
proc print data=aclass3;
run;EXAMPLE 2. Convert all variables names to upcase:
%unifyVarsCaseSize(work,bclass4-bclass6,case=U)
proc print data=bclass4;
proc print data=bclass5;
proc print data=bclass6;
run;EXAMPLE 3. No conversion done:
data work.abc;
abc=42;
run;
%unifyVarsCaseSize(work,abc,debug=1)
proc print data=abc;
run;EXAMPLE 4. Variables in all data sets in WORK converted to upcase:
%unifyVarsCaseSize(work,_ALL_,case=L)
%unifyVarsCaseSize(work,_ALL_,case=U) The workPath() macro function returns path to the WORK library,
it is a wrapper to pathname("work", "L") function.
See examples below for the details.
The %workPath() macro executes like a pure macro code.
The basic syntax is the following, the <...> means optional parameters:
%workPath()Arguments description:
*) No arguments.
EXAMPLE 1. Create new library inside WORK library:
options dlCreateDir;
libname NEW "%workPath()/new";Copyright (c) 2020 - 2024 Bartosz Jablonski
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.