Fix failure of btree_gin indexscans with "char" type and </<= operators.

As a result of confusion about whether the "char" type is signed or
unsigned, scans for index searches like "col < 'x'" or "col <= 'x'"
would start at the middle of the index not the left end, thus missing
many or all of the entries they should find.  Fortunately, this
is not a symptom of index corruption.  It's only the search logic
that is broken, and we can fix it without unpleasant side-effects.

Per report from Jason Kim.  This has been wrong since btree_gin's
beginning, so back-patch to all supported branches.

Discussion: https://postgr.es/m/20210810001649.htnltbh7c63re42p@jasonk.me
This commit is contained in:
Tom Lane 2021-08-10 18:10:29 -04:00
parent 72bbff4cd6
commit a6bd28beb0
2 changed files with 10 additions and 3 deletions

View File

@ -357,7 +357,7 @@ GIN_SUPPORT(bpchar, true, leftmostvalue_text, bpcharcmp)
static Datum
leftmostvalue_char(void)
{
return CharGetDatum(SCHAR_MIN);
return CharGetDatum(0);
}
GIN_SUPPORT(char, false, leftmostvalue_char, btcharcmp)

View File

@ -7,12 +7,19 @@ CREATE INDEX idx_char ON test_char USING gin (i);
SELECT * FROM test_char WHERE i<'d'::"char" ORDER BY i;
i
---
(0 rows)
a
b
c
(3 rows)
SELECT * FROM test_char WHERE i<='d'::"char" ORDER BY i;
i
---
(0 rows)
a
b
c
d
(4 rows)
SELECT * FROM test_char WHERE i='d'::"char" ORDER BY i;
i