Ensure that a tsquery like '!foo' matches empty tsvectors.

!foo means "the tsvector does not contain foo", and therefore it should
match an empty tsvector.  ts_match_vq() overenthusiastically supposed
that an empty tsvector could never match any query, so it forcibly
returned FALSE, the wrong answer.  Remove the premature optimization.

Our behavior on this point was inconsistent, because while seqscans and
GIST index searches both failed to match empty tsvectors, GIN index
searches would find them, since GIN scans don't rely on ts_match_vq().
That makes this certainly a bug, not a debatable definition disagreement,
so back-patch to all supported branches.

Report and diagnosis by Tom Dunstan (bug #14515); added test cases by me.

Discussion: https://postgr.es/m/20170126025524.1434.97828@wrigleys.postgresql.org

diff --git a/src/backend/utils/adt/tsvector_op.c b/src/backend/utils/adt/tsvector_op.c
index 7af4a3bfc18478900eac32c0e0a612250531326a..f1d26fcfa8e9e305b1a52063d0ce0c53efa340e6 100644 (file)
--- a/src/backend/utils/adt/tsvector_op.c
+++ b/src/backend/utils/adt/tsvector_op.c
@@ -1926,7 +1926,8 @@ ts_match_vq(PG_FUNCTION_ARGS)
        CHKVAL          chkval;
        bool            result;
 
-       if (!val->size || !query->size)
+       /* empty query matches nothing */
+       if (!query->size)
        {
                PG_FREE_IF_COPY(val, 0);
                PG_FREE_IF_COPY(query, 1);

diff --git a/src/test/regress/expected/tsearch.out b/src/test/regress/expected/tsearch.out
index 0681d43358bf446be82ea4c1883184c88ae0162b..b2fc9e207edb756a02b756e1fb9621ea3504d4ee 100644 (file)
--- a/src/test/regress/expected/tsearch.out
+++ b/src/test/regress/expected/tsearch.out
@@ -98,8 +98,108 @@ SELECT count(*) FROM test_tsvector WHERE a @@ 'w:*|q:*';
    494
 (1 row)
 
+SELECT count(*) FROM test_tsvector WHERE a @@ any ('{wr,qh}');
+ count 
+-------
+   158
+(1 row)
+
+SELECT count(*) FROM test_tsvector WHERE a @@ 'no_such_lexeme';
+ count 
+-------
+     0
+(1 row)
+
+SELECT count(*) FROM test_tsvector WHERE a @@ '!no_such_lexeme';
+ count 
+-------
+   508
+(1 row)
+
 create index wowidx on test_tsvector using gist (a);
 SET enable_seqscan=OFF;
+SET enable_indexscan=ON;
+SET enable_bitmapscan=OFF;
+explain (costs off) SELECT count(*) FROM test_tsvector WHERE a @@ 'wr|qh';
+                      QUERY PLAN                       
+-------------------------------------------------------
+ Aggregate
+   ->  Index Scan using wowidx on test_tsvector
+         Index Cond: (a @@ '''wr'' | ''qh'''::tsquery)
+(3 rows)
+
+SELECT count(*) FROM test_tsvector WHERE a @@ 'wr|qh';
+ count 
+-------
+   158
+(1 row)
+
+SELECT count(*) FROM test_tsvector WHERE a @@ 'wr&qh';
+ count 
+-------
+    17
+(1 row)
+
+SELECT count(*) FROM test_tsvector WHERE a @@ 'eq&yt';
+ count 
+-------
+     6
+(1 row)
+
+SELECT count(*) FROM test_tsvector WHERE a @@ 'eq|yt';
+ count 
+-------
+    98
+(1 row)
+
+SELECT count(*) FROM test_tsvector WHERE a @@ '(eq&yt)|(wr&qh)';
+ count 
+-------
+    23
+(1 row)
+
+SELECT count(*) FROM test_tsvector WHERE a @@ '(eq|yt)&(wr|qh)';
+ count 
+-------
+    39
+(1 row)
+
+SELECT count(*) FROM test_tsvector WHERE a @@ 'w:*|q:*';
+ count 
+-------
+   494
+(1 row)
+
+SELECT count(*) FROM test_tsvector WHERE a @@ any ('{wr,qh}');
+ count 
+-------
+   158
+(1 row)
+
+SELECT count(*) FROM test_tsvector WHERE a @@ 'no_such_lexeme';
+ count 
+-------
+     0
+(1 row)
+
+SELECT count(*) FROM test_tsvector WHERE a @@ '!no_such_lexeme';
+ count 
+-------
+   508
+(1 row)
+
+SET enable_indexscan=OFF;
+SET enable_bitmapscan=ON;
+explain (costs off) SELECT count(*) FROM test_tsvector WHERE a @@ 'wr|qh';
+                         QUERY PLAN                          
+-------------------------------------------------------------
+ Aggregate
+   ->  Bitmap Heap Scan on test_tsvector
+         Recheck Cond: (a @@ '''wr'' | ''qh'''::tsquery)
+         ->  Bitmap Index Scan on wowidx
+               Index Cond: (a @@ '''wr'' | ''qh'''::tsquery)
+(5 rows)
+
 SELECT count(*) FROM test_tsvector WHERE a @@ 'wr|qh';
  count 
 -------
@@ -148,10 +248,35 @@ SELECT count(*) FROM test_tsvector WHERE a @@ any ('{wr,qh}');
    158
 (1 row)
 
+SELECT count(*) FROM test_tsvector WHERE a @@ 'no_such_lexeme';
+ count 
+-------
+     0
+(1 row)
+
+SELECT count(*) FROM test_tsvector WHERE a @@ '!no_such_lexeme';
+ count 
+-------
+   508
+(1 row)
+
 RESET enable_seqscan;
+RESET enable_indexscan;
+RESET enable_bitmapscan;
 DROP INDEX wowidx;
 CREATE INDEX wowidx ON test_tsvector USING gin (a);
 SET enable_seqscan=OFF;
+-- GIN only supports bitmapscan, so no need to test plain indexscan
+explain (costs off) SELECT count(*) FROM test_tsvector WHERE a @@ 'wr|qh';
+                         QUERY PLAN                          
+-------------------------------------------------------------
+ Aggregate
+   ->  Bitmap Heap Scan on test_tsvector
+         Recheck Cond: (a @@ '''wr'' | ''qh'''::tsquery)
+         ->  Bitmap Index Scan on wowidx
+               Index Cond: (a @@ '''wr'' | ''qh'''::tsquery)
+(5 rows)
+
 SELECT count(*) FROM test_tsvector WHERE a @@ 'wr|qh';
  count 
 -------
@@ -200,6 +325,18 @@ SELECT count(*) FROM test_tsvector WHERE a @@ any ('{wr,qh}');
    158
 (1 row)
 
+SELECT count(*) FROM test_tsvector WHERE a @@ 'no_such_lexeme';
+ count 
+-------
+     0
+(1 row)
+
+SELECT count(*) FROM test_tsvector WHERE a @@ '!no_such_lexeme';
+ count 
+-------
+   508
+(1 row)
+
 RESET enable_seqscan;
 INSERT INTO test_tsvector VALUES ('???', 'DFG:1A,2B,6C,10 FGH');
 SELECT * FROM ts_stat('SELECT a FROM test_tsvector') ORDER BY ndoc DESC, nentry DESC, word LIMIT 10;

diff --git a/src/test/regress/expected/tstypes.out b/src/test/regress/expected/tstypes.out
index dcce82fdc4f99c1abd70ef6b488d06a22d32c0c6..6272e70e09ff5736338b00755e81c8364a517273 100644 (file)
--- a/src/test/regress/expected/tstypes.out
+++ b/src/test/regress/expected/tstypes.out
@@ -773,6 +773,18 @@ select to_tsvector('simple', 'x y q y') @@ '!x <-> y' AS "true";
  t
 (1 row)
 
+select to_tsvector('simple', 'x y q y') @@ '!foo' AS "true";
+ true 
+------
+ t
+(1 row)
+
+select to_tsvector('simple', '') @@ '!foo' AS "true";
+ true 
+------
+ t
+(1 row)
+
 --ranking
 SELECT ts_rank(' a:1 s:2C d g'::tsvector, 'a | s');
   ts_rank  

diff --git a/src/test/regress/sql/tsearch.sql b/src/test/regress/sql/tsearch.sql
index 1255f6954db95498751f08b0b6856fc417af171f..e4b21f8f184ac0833ba8153a786faad587a79bf0 100644 (file)
--- a/src/test/regress/sql/tsearch.sql
+++ b/src/test/regress/sql/tsearch.sql
@@ -48,10 +48,17 @@ SELECT count(*) FROM test_tsvector WHERE a @@ 'eq|yt';
 SELECT count(*) FROM test_tsvector WHERE a @@ '(eq&yt)|(wr&qh)';
 SELECT count(*) FROM test_tsvector WHERE a @@ '(eq|yt)&(wr|qh)';
 SELECT count(*) FROM test_tsvector WHERE a @@ 'w:*|q:*';
+SELECT count(*) FROM test_tsvector WHERE a @@ any ('{wr,qh}');
+SELECT count(*) FROM test_tsvector WHERE a @@ 'no_such_lexeme';
+SELECT count(*) FROM test_tsvector WHERE a @@ '!no_such_lexeme';
 
 create index wowidx on test_tsvector using gist (a);
 
 SET enable_seqscan=OFF;
+SET enable_indexscan=ON;
+SET enable_bitmapscan=OFF;
+
+explain (costs off) SELECT count(*) FROM test_tsvector WHERE a @@ 'wr|qh';
 
 SELECT count(*) FROM test_tsvector WHERE a @@ 'wr|qh';
 SELECT count(*) FROM test_tsvector WHERE a @@ 'wr&qh';
@@ -61,14 +68,37 @@ SELECT count(*) FROM test_tsvector WHERE a @@ '(eq&yt)|(wr&qh)';
 SELECT count(*) FROM test_tsvector WHERE a @@ '(eq|yt)&(wr|qh)';
 SELECT count(*) FROM test_tsvector WHERE a @@ 'w:*|q:*';
 SELECT count(*) FROM test_tsvector WHERE a @@ any ('{wr,qh}');
+SELECT count(*) FROM test_tsvector WHERE a @@ 'no_such_lexeme';
+SELECT count(*) FROM test_tsvector WHERE a @@ '!no_such_lexeme';
+
+SET enable_indexscan=OFF;
+SET enable_bitmapscan=ON;
+
+explain (costs off) SELECT count(*) FROM test_tsvector WHERE a @@ 'wr|qh';
+
+SELECT count(*) FROM test_tsvector WHERE a @@ 'wr|qh';
+SELECT count(*) FROM test_tsvector WHERE a @@ 'wr&qh';
+SELECT count(*) FROM test_tsvector WHERE a @@ 'eq&yt';
+SELECT count(*) FROM test_tsvector WHERE a @@ 'eq|yt';
+SELECT count(*) FROM test_tsvector WHERE a @@ '(eq&yt)|(wr&qh)';
+SELECT count(*) FROM test_tsvector WHERE a @@ '(eq|yt)&(wr|qh)';
+SELECT count(*) FROM test_tsvector WHERE a @@ 'w:*|q:*';
+SELECT count(*) FROM test_tsvector WHERE a @@ any ('{wr,qh}');
+SELECT count(*) FROM test_tsvector WHERE a @@ 'no_such_lexeme';
+SELECT count(*) FROM test_tsvector WHERE a @@ '!no_such_lexeme';
 
 RESET enable_seqscan;
+RESET enable_indexscan;
+RESET enable_bitmapscan;
 
 DROP INDEX wowidx;
 
 CREATE INDEX wowidx ON test_tsvector USING gin (a);
 
 SET enable_seqscan=OFF;
+-- GIN only supports bitmapscan, so no need to test plain indexscan
+
+explain (costs off) SELECT count(*) FROM test_tsvector WHERE a @@ 'wr|qh';
 
 SELECT count(*) FROM test_tsvector WHERE a @@ 'wr|qh';
 SELECT count(*) FROM test_tsvector WHERE a @@ 'wr&qh';
@@ -78,8 +108,11 @@ SELECT count(*) FROM test_tsvector WHERE a @@ '(eq&yt)|(wr&qh)';
 SELECT count(*) FROM test_tsvector WHERE a @@ '(eq|yt)&(wr|qh)';
 SELECT count(*) FROM test_tsvector WHERE a @@ 'w:*|q:*';
 SELECT count(*) FROM test_tsvector WHERE a @@ any ('{wr,qh}');
+SELECT count(*) FROM test_tsvector WHERE a @@ 'no_such_lexeme';
+SELECT count(*) FROM test_tsvector WHERE a @@ '!no_such_lexeme';
 
 RESET enable_seqscan;
+
 INSERT INTO test_tsvector VALUES ('???', 'DFG:1A,2B,6C,10 FGH');
 SELECT * FROM ts_stat('SELECT a FROM test_tsvector') ORDER BY ndoc DESC, nentry DESC, word LIMIT 10;
 SELECT * FROM ts_stat('SELECT a FROM test_tsvector', 'AB') ORDER BY ndoc DESC, nentry DESC, word;

diff --git a/src/test/regress/sql/tstypes.sql b/src/test/regress/sql/tstypes.sql
index 77436ce04e06f88866b4c9855f36559b776dc6b8..0a40ec9350440b67ef2fe2c6cfb9112d0a31be36 100644 (file)
--- a/src/test/regress/sql/tstypes.sql
+++ b/src/test/regress/sql/tstypes.sql
@@ -145,6 +145,8 @@ select to_tsvector('simple', 'x q') @@ '(x | y <-> !z) <-> q' AS "true";
 select to_tsvector('simple', 'x q') @@ '(!x | y <-> z) <-> q' AS "false";
 select to_tsvector('simple', 'z q') @@ '(!x | y <-> z) <-> q' AS "true";
 select to_tsvector('simple', 'x y q y') @@ '!x <-> y' AS "true";
+select to_tsvector('simple', 'x y q y') @@ '!foo' AS "true";
+select to_tsvector('simple', '') @@ '!foo' AS "true";
 
 --ranking
 SELECT ts_rank(' a:1 s:2C d g'::tsvector, 'a | s');