Skip to content

Commit

Permalink
fix: Support double quotes in date_part (apache#10833)
Browse files Browse the repository at this point in the history 
* fix: Support double quotes in date_part.rs

* chore

* chore

* refacor: Use Regex to extract

* chore

* chore: Use replaceen

* chore

* fix
  • Loading branch information
@Weijun-H
Weijun-H authored Jun 10, 2024
1 parent f6450e2 commit e094f94
Show file tree
Hide file tree
Showing 2 changed files with 117 additions and 1 deletion.
8 changes: 7 additions & 1 deletion datafusion/functions/src/datetime/date_part.rs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -127,7 +127,13 @@ impl ScalarUDFImpl for DatePartFunc {
ColumnarValue::Scalar(scalar) => scalar.to_array()?,
};

let arr = match part.to_lowercase().as_str() {
// to remove quotes at most 2 characters
let part_trim = part.trim_matches(|c| c == '\'' || c == '\"');
if ![2, 0].contains(&(part.len() - part_trim.len())) {
return exec_err!("Date part '{part}' not supported");
}

let arr = match part_trim.to_lowercase().as_str() {
"year" => date_part_f64(array.as_ref(), DatePart::Year)?,
"quarter" => date_part_f64(array.as_ref(), DatePart::Quarter)?,
"month" => date_part_f64(array.as_ref(), DatePart::Month)?,
Expand Down
110 changes: 110 additions & 0 deletions datafusion/sqllogictest/test_files/expr.slt
View file
Original file line number Diff line number Diff line change
Expand Up @@ -827,6 +827,12 @@ SELECT

# test_extract_date_part

query error
SELECT EXTRACT("'''year'''" FROM timestamp '2020-09-08T12:00:00+00:00')

query error
SELECT EXTRACT("'year'" FROM timestamp '2020-09-08T12:00:00+00:00')

query R
SELECT date_part('YEAR', CAST('2000-01-01' AS DATE))
----
Expand All @@ -837,6 +843,14 @@ SELECT EXTRACT(year FROM timestamp '2020-09-08T12:00:00+00:00')
----
2020

query R
SELECT EXTRACT("year" FROM timestamp '2020-09-08T12:00:00+00:00')
----
2020

query error
SELECT EXTRACT('year' FROM timestamp '2020-09-08T12:00:00+00:00')

query R
SELECT date_part('QUARTER', CAST('2000-01-01' AS DATE))
----
Expand All @@ -847,6 +861,14 @@ SELECT EXTRACT(quarter FROM to_timestamp('2020-09-08T12:00:00+00:00'))
----
3

query R
SELECT EXTRACT("quarter" FROM to_timestamp('2020-09-08T12:00:00+00:00'))
----
3

query error
SELECT EXTRACT('quarter' FROM to_timestamp('2020-09-08T12:00:00+00:00'))

query R
SELECT date_part('MONTH', CAST('2000-01-01' AS DATE))
----
Expand All @@ -857,6 +879,14 @@ SELECT EXTRACT(month FROM to_timestamp('2020-09-08T12:00:00+00:00'))
----
9

query R
SELECT EXTRACT("month" FROM to_timestamp('2020-09-08T12:00:00+00:00'))
----
9

query error
SELECT EXTRACT('month' FROM to_timestamp('2020-09-08T12:00:00+00:00'))

query R
SELECT date_part('WEEK', CAST('2003-01-01' AS DATE))
----
Expand All @@ -867,6 +897,14 @@ SELECT EXTRACT(WEEK FROM to_timestamp('2020-09-08T12:00:00+00:00'))
----
37

query R
SELECT EXTRACT("WEEK" FROM to_timestamp('2020-09-08T12:00:00+00:00'))
----
37

query error
SELECT EXTRACT('WEEK' FROM to_timestamp('2020-09-08T12:00:00+00:00'))

query R
SELECT date_part('DAY', CAST('2000-01-01' AS DATE))
----
Expand All @@ -877,6 +915,14 @@ SELECT EXTRACT(day FROM to_timestamp('2020-09-08T12:00:00+00:00'))
----
8

query R
SELECT EXTRACT("day" FROM to_timestamp('2020-09-08T12:00:00+00:00'))
----
8

query error
SELECT EXTRACT('day' FROM to_timestamp('2020-09-08T12:00:00+00:00'))

query R
SELECT date_part('DOY', CAST('2000-01-01' AS DATE))
----
Expand All @@ -887,6 +933,14 @@ SELECT EXTRACT(doy FROM to_timestamp('2020-09-08T12:00:00+00:00'))
----
252

query R
SELECT EXTRACT("doy" FROM to_timestamp('2020-09-08T12:00:00+00:00'))
----
252

query error
SELECT EXTRACT('doy' FROM to_timestamp('2020-09-08T12:00:00+00:00'))

query R
SELECT date_part('DOW', CAST('2000-01-01' AS DATE))
----
Expand All @@ -897,6 +951,14 @@ SELECT EXTRACT(dow FROM to_timestamp('2020-09-08T12:00:00+00:00'))
----
2

query R
SELECT EXTRACT("dow" FROM to_timestamp('2020-09-08T12:00:00+00:00'))
----
2

query error
SELECT EXTRACT('dow' FROM to_timestamp('2020-09-08T12:00:00+00:00'))

query R
SELECT date_part('HOUR', CAST('2000-01-01' AS DATE))
----
Expand All @@ -907,11 +969,27 @@ SELECT EXTRACT(hour FROM to_timestamp('2020-09-08T12:03:03+00:00'))
----
12

query R
SELECT EXTRACT("hour" FROM to_timestamp('2020-09-08T12:03:03+00:00'))
----
12

query error
SELECT EXTRACT('hour' FROM to_timestamp('2020-09-08T12:03:03+00:00'))

query R
SELECT EXTRACT(minute FROM to_timestamp('2020-09-08T12:12:00+00:00'))
----
12

query R
SELECT EXTRACT("minute" FROM to_timestamp('2020-09-08T12:12:00+00:00'))
----
12

query error
SELECT EXTRACT('minute' FROM to_timestamp('2020-09-08T12:12:00+00:00'))

query R
SELECT date_part('minute', to_timestamp('2020-09-08T12:12:00+00:00'))
----
Expand All @@ -937,6 +1015,38 @@ SELECT EXTRACT(nanosecond FROM timestamp '2020-09-08T12:00:12.12345678+00:00')
----
12123456780

query R
SELECT EXTRACT("second" FROM timestamp '2020-09-08T12:00:12.12345678+00:00')
----
12.12345678

query R
SELECT EXTRACT("millisecond" FROM timestamp '2020-09-08T12:00:12.12345678+00:00')
----
12123.45678

query R
SELECT EXTRACT("microsecond" FROM timestamp '2020-09-08T12:00:12.12345678+00:00')
----
12123456.78

query R
SELECT EXTRACT("nanosecond" FROM timestamp '2020-09-08T12:00:12.12345678+00:00')
----
12123456780

query error
SELECT EXTRACT('second' FROM timestamp '2020-09-08T12:00:12.12345678+00:00')

query error
SELECT EXTRACT('millisecond' FROM timestamp '2020-09-08T12:00:12.12345678+00:00')

query error
SELECT EXTRACT('microsecond' FROM timestamp '2020-09-08T12:00:12.12345678+00:00')

query error
SELECT EXTRACT('nanosecond' FROM timestamp '2020-09-08T12:00:12.12345678+00:00')

# Keep precision when coercing Utf8 to Timestamp
query R
SELECT date_part('second', timestamp '2020-09-08T12:00:12.12345678+00:00')
Expand Down

0 comments on commit e094f94

Please sign in to comment.