summaryrefslogtreecommitdiff
path: root/src/test
diff options
context:
space:
mode:
authorTom Lane <tgl@sss.pgh.pa.us>2023-12-27 15:55:46 -0500
committerTom Lane <tgl@sss.pgh.pa.us>2023-12-27 15:55:46 -0500
commit58054de2d0847c09ef091956f72ae5e9fb9a176e (patch)
tree509bfa191ff6a4c455b8b7ac269d9acdcb29d4f3 /src/test
parent6c361d323b908c0391e05d15afc01957db2e415b (diff)
Improve the implementation of information_schema._pg_expandarray().
This function was originally coded with a handmade expansion of the array subscripts. We can do it a little faster and far more legibly today, by using unnest() WITH ORDINALITY. While at it, let's apply the rowcount estimation support that exists for the underlying unnest() function: reduce the default ROWS estimate to 100 and attach array_unnest_support. I'm not sure that array_unnest_support can do anything useful today with the call sites that exist in information_schema, but it can't hurt, and the existing default rowcount of 1000 is surely much too high for any of these cases. The psql.sql regression script is using _pg_expandarray() as a test case for \sf+. While we could keep doing so, the new one-line function body makes a poor test case for \sf+ row-numbering, so switch it to print another information_schema function. Discussion: https://postgr.es/m/1424303.1703355485@sss.pgh.pa.us
Diffstat (limited to 'src/test')
-rw-r--r--src/test/regress/expected/psql.out40
-rw-r--r--src/test/regress/sql/psql.sql4
2 files changed, 24 insertions, 20 deletions
diff --git a/src/test/regress/expected/psql.out b/src/test/regress/expected/psql.out
index 13e4f6db7ba..5d61e4c7bb9 100644
--- a/src/test/regress/expected/psql.out
+++ b/src/test/regress/expected/psql.out
@@ -5293,26 +5293,30 @@ comment on function psql_df_plpgsql () is 'some comment';
rollback;
drop role regress_psql_user;
-- check \sf
-\sf information_schema._pg_expandarray
-CREATE OR REPLACE FUNCTION information_schema._pg_expandarray(anyarray, OUT x anyelement, OUT n integer)
- RETURNS SETOF record
+\sf information_schema._pg_index_position
+CREATE OR REPLACE FUNCTION information_schema._pg_index_position(oid, smallint)
+ RETURNS integer
LANGUAGE sql
- IMMUTABLE PARALLEL SAFE STRICT
-AS $function$select $1[s],
- s operator(pg_catalog.-) pg_catalog.array_lower($1,1) operator(pg_catalog.+) 1
- from pg_catalog.generate_series(pg_catalog.array_lower($1,1),
- pg_catalog.array_upper($1,1),
- 1) as g(s)$function$
-\sf+ information_schema._pg_expandarray
- CREATE OR REPLACE FUNCTION information_schema._pg_expandarray(anyarray, OUT x anyelement, OUT n integer)
- RETURNS SETOF record
+ STABLE STRICT
+BEGIN ATOMIC
+ SELECT (ss.a).n AS n
+ FROM ( SELECT information_schema._pg_expandarray(pg_index.indkey) AS a
+ FROM pg_index
+ WHERE (pg_index.indexrelid = $1)) ss
+ WHERE ((ss.a).x = $2);
+END
+\sf+ information_schema._pg_index_position
+ CREATE OR REPLACE FUNCTION information_schema._pg_index_position(oid, smallint)
+ RETURNS integer
LANGUAGE sql
- IMMUTABLE PARALLEL SAFE STRICT
-1 AS $function$select $1[s],
-2 s operator(pg_catalog.-) pg_catalog.array_lower($1,1) operator(pg_catalog.+) 1
-3 from pg_catalog.generate_series(pg_catalog.array_lower($1,1),
-4 pg_catalog.array_upper($1,1),
-5 1) as g(s)$function$
+ STABLE STRICT
+1 BEGIN ATOMIC
+2 SELECT (ss.a).n AS n
+3 FROM ( SELECT information_schema._pg_expandarray(pg_index.indkey) AS a
+4 FROM pg_index
+5 WHERE (pg_index.indexrelid = $1)) ss
+6 WHERE ((ss.a).x = $2);
+7 END
\sf+ interval_pl_time
CREATE OR REPLACE FUNCTION pg_catalog.interval_pl_time(interval, time without time zone)
RETURNS time without time zone
diff --git a/src/test/regress/sql/psql.sql b/src/test/regress/sql/psql.sql
index 695c72d8668..f199d624d3b 100644
--- a/src/test/regress/sql/psql.sql
+++ b/src/test/regress/sql/psql.sql
@@ -1312,8 +1312,8 @@ rollback;
drop role regress_psql_user;
-- check \sf
-\sf information_schema._pg_expandarray
-\sf+ information_schema._pg_expandarray
+\sf information_schema._pg_index_position
+\sf+ information_schema._pg_index_position
\sf+ interval_pl_time
\sf ts_debug(text)
\sf+ ts_debug(text)