Extend EXPLAIN and add config param to switch transformation of prope…

…rty filter (apache#1262) * Add explain options Explain command in the following format is supported now: `EXPLAIN (VERBOSE, COSTS OFF, FORMAT XML) ...` Note that, this is basically Postgres' EXPLAIN command, and the purpose of this is to support debugging and regression tests. * Add config param to switch transformation method of property filter When the `age.enable_containment` parameter is on, the agtype containment operator is used to transform property filter. When off, access operator is used instead. The former case is preferable for GIN index, and the later for BTREE expression index. The idea of replacing containment with access operator in order to support BTREE index is taken from a patch by Josh Innis. A note on regression testing- although there are test cases for the `age.enable_containment` parameter, sometimes it may be useful to set this parameter before running any tests (not just the ones related to it). For example, when the logic related to property transformation changes. It can be set in the `age_regression.conf` file.
Zainab-Saad · Mar 6, 2024 · 3bef830 · 3bef830
1 parent 30d9399
commit 3bef830
Show file tree

Hide file tree

Showing 9 changed files with 572 additions and 8 deletions.
diff --git a/Makefile b/Makefile
@@ -69,7 +69,8 @@ OBJS = src/backend/age.o \
        src/backend/utils/load/ag_load_edges.o \
        src/backend/utils/load/age_load.o \
        src/backend/utils/load/libcsv.o \
-       src/backend/utils/name_validation.o
+       src/backend/utils/name_validation.o \
+       src/backend/utils/ag_guc.o
 
 EXTENSION = age
 
@@ -105,7 +106,7 @@ REGRESS = scan \
 srcdir=`pwd`
 
 ag_regress_dir = $(srcdir)/regress
-REGRESS_OPTS = --load-extension=age --inputdir=$(ag_regress_dir) --outputdir=$(ag_regress_dir) --temp-instance=$(ag_regress_dir)/instance --port=61958 --encoding=UTF-8
+REGRESS_OPTS = --load-extension=age --inputdir=$(ag_regress_dir) --outputdir=$(ag_regress_dir) --temp-instance=$(ag_regress_dir)/instance --port=61958 --encoding=UTF-8 --temp-config $(ag_regress_dir)/age_regression.conf
 
 ag_regress_out = instance/ log/ results/ regression.*
 EXTRA_CLEAN = $(addprefix $(ag_regress_dir)/, $(ag_regress_out)) src/backend/parser/cypher_gram.c src/include/parser/cypher_gram_def.h src/include/parser/cypher_kwlist_d.h

diff --git a/regress/age_regression.conf b/regress/age_regression.conf
@@ -0,0 +1 @@
+#age.enable_containment = on
diff --git a/regress/expected/cypher_match.out b/regress/expected/cypher_match.out
@@ -2419,6 +2419,170 @@ SELECT * FROM cypher('cypher_match', $$ MATCH p=(x:r)-[]->(x:R) RETURN p, x $$)
 ERROR:  multiple labels for variable 'x' are not supported
 LINE 1: ...FROM cypher('cypher_match', $$ MATCH p=(x:r)-[]->(x:R) RETUR...
                                                              ^
+--
+-- Test age.enable_containment configuration parameter
+--
+-- Test queries are run before and after switching off this parameter.
+-- When  on, the containment operator should be used to filter properties.
+-- When off, the access operator should be used.
+--
+SELECT create_graph('test_enable_containment');
+NOTICE:  graph "test_enable_containment" has been created
+ create_graph 
+--------------
+
+(1 row)
+
+SELECT * FROM cypher('test_enable_containment',
+$$
+    CREATE (x:Customer {
+        name: 'Bob',
+        school: {
+            name: 'XYZ College',
+            program: {
+                major: 'Psyc',
+                degree: 'BSc'
+            }
+        },
+        phone: [ 123456789, 987654321, 456987123 ],
+        addr: [
+            {city: 'Vancouver', street: 30},
+            {city: 'Toronto', street: 40}
+        ]
+    })
+    RETURN x
+$$) as (a agtype);
+                                                                                                                                                  a                                                                                                                                                  
+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
+ {"id": 844424930131969, "label": "Customer", "properties": {"addr": [{"city": "Vancouver", "street": 30}, {"city": "Toronto", "street": 40}], "name": "Bob", "phone": [123456789, 987654321, 456987123], "school": {"name": "XYZ College", "program": {"major": "Psyc", "degree": "BSc"}}}}::vertex
+(1 row)
+
+-- With enable_containment on
+SET age.enable_containment = on;
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {addr:[{city:'Toronto'}]}) RETURN x $$) as (a agtype);
+ count 
+-------
+     1
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {addr:[{city:'Toronto'}, {city: 'Vancouver'}]}) RETURN x $$) as (a agtype);
+ count 
+-------
+     1
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {addr:[{city:'Alberta'}]}) RETURN x $$) as (a agtype);
+ count 
+-------
+     0
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {school:{program:{major:'Psyc'}}}) RETURN x $$) as (a agtype);
+ count 
+-------
+     1
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {name:'Bob',school:{program:{degree:'BSc'}}}) RETURN x $$) as (a agtype);
+ count 
+-------
+     1
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {school:{program:{major:'Cs'}}}) RETURN x $$) as (a agtype);
+ count 
+-------
+     0
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {name:'Bob',school:{program:{degree:'PHd'}}}) RETURN x $$) as (a agtype);
+ count 
+-------
+     0
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {phone:[987654321]}) RETURN x $$) as (a agtype);
+ count 
+-------
+     1
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {phone:[654765876]}) RETURN x $$) as (a agtype);
+ count 
+-------
+     0
+(1 row)
+
+SELECT * FROM cypher('test_enable_containment', $$ EXPLAIN (COSTS OFF) MATCH (x:Customer {school:{name:'XYZ',program:{degree:'BSc'}},phone:[987654321],parents:{}}) RETURN x $$) as (a agtype);
+                                                             QUERY PLAN                                                             
+------------------------------------------------------------------------------------------------------------------------------------
+ Seq Scan on "Customer" x
+   Filter: (properties @> '{"phone": [987654321], "school": {"name": "XYZ", "program": {"degree": "BSc"}}, "parents": {}}'::agtype)
+(2 rows)
+
+-- Previous set of queries, with enable_containment off
+SET age.enable_containment = off;
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {addr:[{city:'Toronto'}]}) RETURN x $$) as (a agtype);
+ count 
+-------
+     1
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {addr:[{city:'Toronto'}, {city: 'Vancouver'}]}) RETURN x $$) as (a agtype);
+ count 
+-------
+     1
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {addr:[{city:'Alberta'}]}) RETURN x $$) as (a agtype);
+ count 
+-------
+     0
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {school:{program:{major:'Psyc'}}}) RETURN x $$) as (a agtype);
+ count 
+-------
+     1
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {name:'Bob',school:{program:{degree:'BSc'}}}) RETURN x $$) as (a agtype);
+ count 
+-------
+     1
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {school:{program:{major:'Cs'}}}) RETURN x $$) as (a agtype);
+ count 
+-------
+     0
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {name:'Bob',school:{program:{degree:'PHd'}}}) RETURN x $$) as (a agtype);
+ count 
+-------
+     0
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {phone:[987654321]}) RETURN x $$) as (a agtype);
+ count 
+-------
+     1
+(1 row)
+
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {phone:[654765876]}) RETURN x $$) as (a agtype);
+ count 
+-------
+     0
+(1 row)
+
+SELECT * FROM cypher('test_enable_containment', $$ EXPLAIN (COSTS OFF) MATCH (x:Customer {school:{name:'XYZ',program:{degree:'BSc'}},phone:[987654321],parents:{}}) RETURN x $$) as (a agtype);
+                                                                                                                                                                                                                             QUERY PLAN                                                                                                                                                                                                                              
+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
+ Seq Scan on "Customer" x
+   Filter: ((agtype_access_operator(VARIADIC ARRAY[properties, '"school"'::agtype, '"name"'::agtype]) = '"XYZ"'::agtype) AND (agtype_access_operator(VARIADIC ARRAY[properties, '"school"'::agtype, '"program"'::agtype, '"degree"'::agtype]) = '"BSc"'::agtype) AND (agtype_access_operator(VARIADIC ARRAY[properties, '"phone"'::agtype]) @> '[987654321]'::agtype) AND (agtype_access_operator(VARIADIC ARRAY[properties, '"parents"'::agtype]) @> '{}'::agtype))
+(2 rows)
+
 --
 -- Clean up
 --
@@ -2460,6 +2624,17 @@ NOTICE:  graph "test_retrieve_var" has been dropped
 
 (1 row)
 
+SELECT drop_graph('test_enable_containment', true);
+NOTICE:  drop cascades to 3 other objects
+DETAIL:  drop cascades to table test_enable_containment._ag_label_vertex
+drop cascades to table test_enable_containment._ag_label_edge
+drop cascades to table test_enable_containment."Customer"
+NOTICE:  graph "test_enable_containment" has been dropped
+ drop_graph 
+------------
+
+(1 row)
+
 --
 -- End
 --
diff --git a/regress/sql/cypher_match.sql b/regress/sql/cypher_match.sql
@@ -1077,11 +1077,67 @@ SELECT * FROM cypher('cypher_match', $$ CREATE () WITH * MATCH (x{n0:x.n1}) RETU
 SELECT * FROM cypher('cypher_match', $$ MATCH p=(x)-[]->(x:R) RETURN p, x $$) AS (p agtype, x agtype);
 SELECT * FROM cypher('cypher_match', $$ MATCH p=(x:r)-[]->(x:R) RETURN p, x $$) AS (p agtype, x agtype);
 
+--
+-- Test age.enable_containment configuration parameter
+--
+-- Test queries are run before and after switching off this parameter.
+-- When  on, the containment operator should be used to filter properties.
+-- When off, the access operator should be used.
+--
+
+SELECT create_graph('test_enable_containment');
+SELECT * FROM cypher('test_enable_containment',
+$$
+    CREATE (x:Customer {
+        name: 'Bob',
+        school: {
+            name: 'XYZ College',
+            program: {
+                major: 'Psyc',
+                degree: 'BSc'
+            }
+        },
+        phone: [ 123456789, 987654321, 456987123 ],
+        addr: [
+            {city: 'Vancouver', street: 30},
+            {city: 'Toronto', street: 40}
+        ]
+    })
+    RETURN x
+$$) as (a agtype);
+
+-- With enable_containment on
+SET age.enable_containment = on;
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {addr:[{city:'Toronto'}]}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {addr:[{city:'Toronto'}, {city: 'Vancouver'}]}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {addr:[{city:'Alberta'}]}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {school:{program:{major:'Psyc'}}}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {name:'Bob',school:{program:{degree:'BSc'}}}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {school:{program:{major:'Cs'}}}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {name:'Bob',school:{program:{degree:'PHd'}}}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {phone:[987654321]}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {phone:[654765876]}) RETURN x $$) as (a agtype);
+SELECT * FROM cypher('test_enable_containment', $$ EXPLAIN (COSTS OFF) MATCH (x:Customer {school:{name:'XYZ',program:{degree:'BSc'}},phone:[987654321],parents:{}}) RETURN x $$) as (a agtype);
+
+-- Previous set of queries, with enable_containment off
+SET age.enable_containment = off;
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {addr:[{city:'Toronto'}]}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {addr:[{city:'Toronto'}, {city: 'Vancouver'}]}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {addr:[{city:'Alberta'}]}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {school:{program:{major:'Psyc'}}}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {name:'Bob',school:{program:{degree:'BSc'}}}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {school:{program:{major:'Cs'}}}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {name:'Bob',school:{program:{degree:'PHd'}}}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {phone:[987654321]}) RETURN x $$) as (a agtype);
+SELECT count(*) FROM cypher('test_enable_containment', $$ MATCH (x:Customer {phone:[654765876]}) RETURN x $$) as (a agtype);
+SELECT * FROM cypher('test_enable_containment', $$ EXPLAIN (COSTS OFF) MATCH (x:Customer {school:{name:'XYZ',program:{degree:'BSc'}},phone:[987654321],parents:{}}) RETURN x $$) as (a agtype);
+
 --
 -- Clean up
 --
 SELECT drop_graph('cypher_match', true);
 SELECT drop_graph('test_retrieve_var', true);
+SELECT drop_graph('test_enable_containment', true);
 
 --
 -- End

diff --git a/src/backend/age.c b/src/backend/age.c
@@ -25,6 +25,7 @@
 #include "nodes/ag_nodes.h"
 #include "optimizer/cypher_paths.h"
 #include "parser/cypher_analyze.h"
+#include "utils/ag_guc.h"
 
 PG_MODULE_MAGIC;
 
@@ -37,6 +38,7 @@ void _PG_init(void)
     object_access_hook_init();
     process_utility_hook_init();
     post_parse_analyze_init();
+    define_config_params();
 }
 
 void _PG_fini(void);