Alter the xxx_pattern_ops opclasses to use the regular equality operator of

the associated datatype as their equality member.  This means that these
opclasses can now support plain equality comparisons along with LIKE tests,
thus avoiding the need for an extra index in some applications.  This
optimization was not possible when the pattern opclasses were first introduced,
because we didn't insist that text equality meant bitwise equality; but we
do now, so there is no semantic difference between regular and pattern
equality operators.

I removed the name_pattern_ops opclass altogether, since it's really useless:
name's regular comparisons are just strcmp() and are unlikely to become
something different.  Instead teach indxpath.c that btree name_ops can be
used for LIKE whether or not the locale is C.  This might lead to a useful
speedup in LIKE queries on the system catalogs in non-C locales.

The ~=~ and ~<>~ operators are gone altogether.  (It would have been nice to
keep them for backward compatibility's sake, but since the pg_amop structure
doesn't allow multiple equality operators per opclass, there's no way.)

A not-immediately-obvious incompatibility is that the sort order within
bpchar_pattern_ops indexes changes --- it had been identical to plain
strcmp, but is now trailing-blank-insensitive.  This will impact
in-place upgrades, if those ever happen.

Per discussions a couple months ago.

Author: tglsfdc

doc/src/sgml/indices.sgml

@@ -1,4 +1,4 @@
-<!-- $PostgreSQL: pgsql/doc/src/sgml/indices.sgml,v 1.72 2007/11/13 23:36:26 tgl Exp $ -->
+<!-- $PostgreSQL: pgsql/doc/src/sgml/indices.sgml,v 1.73 2008/05/27 00:13:08 tgl Exp $ -->
 
 <chapter id="indexes">
  <title id="indexes-title">Indexes</title>
@@ -906,11 +906,10 @@ CREATE INDEX <replaceable>name</replaceable> ON <replaceable>table</replaceable>
     <listitem>
      <para>
       The operator classes <literal>text_pattern_ops</literal>,
-      <literal>varchar_pattern_ops</literal>,
-      <literal>bpchar_pattern_ops</literal>, and
-      <literal>name_pattern_ops</literal> support B-tree indexes on
-      the types <type>text</type>, <type>varchar</type>,
-      <type>char</type>, and <type>name</type>, respectively.  The
+      <literal>varchar_pattern_ops</literal>, and
+      <literal>bpchar_pattern_ops</literal> support B-tree indexes on
+      the types <type>text</type>, <type>varchar</type>, and
+      <type>char</type> respectively.  The
       difference from the default operator classes is that the values
       are compared strictly character by character rather than
       according to the locale-specific collation rules.  This makes
@@ -923,10 +922,12 @@ CREATE INDEX <replaceable>name</replaceable> ON <replaceable>table</replaceable>
 CREATE INDEX test_index ON test_table (col varchar_pattern_ops);
 </programlisting>
       Note that you should also create an index with the default operator
-      class if you want queries involving ordinary comparisons to use an
-      index.  Such queries cannot use the
+      class if you want queries involving ordinary <literal>&lt;</>,
+      <literal>&lt;=</>, <literal>&gt;</>, or <literal>&gt;=</> comparisons
+      to use an index.  Such queries cannot use the
       <literal><replaceable>xxx</replaceable>_pattern_ops</literal>
-      operator classes.  It is allowed to create multiple
+      operator classes.  (Ordinary equality comparisons can use these
+      operator classes, however.)  It is allowed to create multiple
       indexes on the same column with different operator classes.
       If you do use the C locale, you do not need the
       <literal><replaceable>xxx</replaceable>_pattern_ops</literal>

src/backend/access/nbtree/nbtcompare.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/access/nbtree/nbtcompare.c,v 1.55 2008/01/01 19:45:46 momjian Exp $
+ *	  $PostgreSQL: pgsql/src/backend/access/nbtree/nbtcompare.c,v 1.56 2008/05/27 00:13:08 tgl Exp $
  *
  * NOTES
  *
@@ -237,12 +237,3 @@ btnamecmp(PG_FUNCTION_ARGS)
 
 	PG_RETURN_INT32(strncmp(NameStr(*a), NameStr(*b), NAMEDATALEN));
 }
-
-Datum
-btname_pattern_cmp(PG_FUNCTION_ARGS)
-{
-	Name		a = PG_GETARG_NAME(0);
-	Name		b = PG_GETARG_NAME(1);
-
-	PG_RETURN_INT32(memcmp(NameStr(*a), NameStr(*b), NAMEDATALEN));
-}

src/backend/optimizer/path/indxpath.c

@@ -9,7 +9,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/optimizer/path/indxpath.c,v 1.230 2008/05/16 16:31:01 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/optimizer/path/indxpath.c,v 1.231 2008/05/27 00:13:09 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -2178,9 +2178,8 @@ match_special_index_operator(Expr *clause, Oid opfamily,
 		case OID_NAME_ICLIKE_OP:
 		case OID_NAME_REGEXEQ_OP:
 		case OID_NAME_ICREGEXEQ_OP:
-			isIndexable =
-				(opfamily == NAME_PATTERN_BTREE_FAM_OID) ||
-				(opfamily == NAME_BTREE_FAM_OID && lc_collate_is_c());
+			/* name uses locale-insensitive sorting */
+			isIndexable = (opfamily == NAME_BTREE_FAM_OID);
 			break;
 
 		case OID_BYTEA_LIKE_OP:
@@ -2700,7 +2699,6 @@ prefix_quals(Node *leftop, Oid opfamily,
 			break;
 
 		case NAME_BTREE_FAM_OID:
-		case NAME_PATTERN_BTREE_FAM_OID:
 			datatype = NAMEOID;
 			break;
 

src/backend/utils/adt/name.c

@@ -14,7 +14,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/utils/adt/name.c,v 1.61 2008/01/01 19:45:52 momjian Exp $
+ *	  $PostgreSQL: pgsql/src/backend/utils/adt/name.c,v 1.62 2008/05/27 00:13:09 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -181,65 +181,6 @@ namege(PG_FUNCTION_ARGS)
 }
 
 
-/*
- * comparison routines for LIKE indexing support
- */
-
-Datum
-name_pattern_eq(PG_FUNCTION_ARGS)
-{
-	Name		arg1 = PG_GETARG_NAME(0);
-	Name		arg2 = PG_GETARG_NAME(1);
-
-	PG_RETURN_BOOL(memcmp(NameStr(*arg1), NameStr(*arg2), NAMEDATALEN) == 0);
-}
-
-Datum
-name_pattern_ne(PG_FUNCTION_ARGS)
-{
-	Name		arg1 = PG_GETARG_NAME(0);
-	Name		arg2 = PG_GETARG_NAME(1);
-
-	PG_RETURN_BOOL(memcmp(NameStr(*arg1), NameStr(*arg2), NAMEDATALEN) != 0);
-}
-
-Datum
-name_pattern_lt(PG_FUNCTION_ARGS)
-{
-	Name		arg1 = PG_GETARG_NAME(0);
-	Name		arg2 = PG_GETARG_NAME(1);
-
-	PG_RETURN_BOOL(memcmp(NameStr(*arg1), NameStr(*arg2), NAMEDATALEN) < 0);
-}
-
-Datum
-name_pattern_le(PG_FUNCTION_ARGS)
-{
-	Name		arg1 = PG_GETARG_NAME(0);
-	Name		arg2 = PG_GETARG_NAME(1);
-
-	PG_RETURN_BOOL(memcmp(NameStr(*arg1), NameStr(*arg2), NAMEDATALEN) <= 0);
-}
-
-Datum
-name_pattern_gt(PG_FUNCTION_ARGS)
-{
-	Name		arg1 = PG_GETARG_NAME(0);
-	Name		arg2 = PG_GETARG_NAME(1);
-
-	PG_RETURN_BOOL(memcmp(NameStr(*arg1), NameStr(*arg2), NAMEDATALEN) > 0);
-}
-
-Datum
-name_pattern_ge(PG_FUNCTION_ARGS)
-{
-	Name		arg1 = PG_GETARG_NAME(0);
-	Name		arg2 = PG_GETARG_NAME(1);
-
-	PG_RETURN_BOOL(memcmp(NameStr(*arg1), NameStr(*arg2), NAMEDATALEN) >= 0);
-}
-
-
 /* (see char.c for comparison/operation routines) */
 
 int

src/backend/utils/adt/varchar.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/utils/adt/varchar.c,v 1.128 2008/05/04 16:42:41 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/utils/adt/varchar.c,v 1.129 2008/05/27 00:13:09 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -886,3 +886,112 @@ hashbpchar(PG_FUNCTION_ARGS)
 
 	return result;
 }
+
+
+/*
+ * The following operators support character-by-character comparison
+ * of bpchar datums, to allow building indexes suitable for LIKE clauses.
+ * Note that the regular bpchareq/bpcharne comparison operators are assumed
+ * to be compatible with these!
+ */
+
+static int
+internal_bpchar_pattern_compare(BpChar *arg1, BpChar *arg2)
+{
+	int			result;
+	int			len1,
+				len2;
+
+	len1 = bcTruelen(arg1);
+	len2 = bcTruelen(arg2);
+
+	result = strncmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2));
+	if (result != 0)
+		return result;
+	else if (len1 < len2)
+		return -1;
+	else if (len1 > len2)
+		return 1;
+	else
+		return 0;
+}
+
+
+Datum
+bpchar_pattern_lt(PG_FUNCTION_ARGS)
+{
+	BpChar	   *arg1 = PG_GETARG_BPCHAR_PP(0);
+	BpChar	   *arg2 = PG_GETARG_BPCHAR_PP(1);
+	int			result;
+
+	result = internal_bpchar_pattern_compare(arg1, arg2);
+
+	PG_FREE_IF_COPY(arg1, 0);
+	PG_FREE_IF_COPY(arg2, 1);
+
+	PG_RETURN_BOOL(result < 0);
+}
+
+
+Datum
+bpchar_pattern_le(PG_FUNCTION_ARGS)
+{
+	BpChar	   *arg1 = PG_GETARG_BPCHAR_PP(0);
+	BpChar	   *arg2 = PG_GETARG_BPCHAR_PP(1);
+	int			result;
+
+	result = internal_bpchar_pattern_compare(arg1, arg2);
+
+	PG_FREE_IF_COPY(arg1, 0);
+	PG_FREE_IF_COPY(arg2, 1);
+
+	PG_RETURN_BOOL(result <= 0);
+}
+
+
+Datum
+bpchar_pattern_ge(PG_FUNCTION_ARGS)
+{
+	BpChar	   *arg1 = PG_GETARG_BPCHAR_PP(0);
+	BpChar	   *arg2 = PG_GETARG_BPCHAR_PP(1);
+	int			result;
+
+	result = internal_bpchar_pattern_compare(arg1, arg2);
+
+	PG_FREE_IF_COPY(arg1, 0);
+	PG_FREE_IF_COPY(arg2, 1);
+
+	PG_RETURN_BOOL(result >= 0);
+}
+
+
+Datum
+bpchar_pattern_gt(PG_FUNCTION_ARGS)
+{
+	BpChar	   *arg1 = PG_GETARG_BPCHAR_PP(0);
+	BpChar	   *arg2 = PG_GETARG_BPCHAR_PP(1);
+	int			result;
+
+	result = internal_bpchar_pattern_compare(arg1, arg2);
+
+	PG_FREE_IF_COPY(arg1, 0);
+	PG_FREE_IF_COPY(arg2, 1);
+
+	PG_RETURN_BOOL(result > 0);
+}
+
+
+Datum
+btbpchar_pattern_cmp(PG_FUNCTION_ARGS)
+{
+	BpChar	   *arg1 = PG_GETARG_BPCHAR_PP(0);
+	BpChar	   *arg2 = PG_GETARG_BPCHAR_PP(1);
+	int			result;
+
+	result = internal_bpchar_pattern_compare(arg1, arg2);
+
+	PG_FREE_IF_COPY(arg1, 0);
+	PG_FREE_IF_COPY(arg2, 1);
+
+	PG_RETURN_INT32(result);
+}

src/backend/utils/adt/varlena.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/utils/adt/varlena.c,v 1.166 2008/05/12 00:00:51 alvherre Exp $
+ *	  $PostgreSQL: pgsql/src/backend/utils/adt/varlena.c,v 1.167 2008/05/27 00:13:09 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -1309,22 +1309,27 @@ text_smaller(PG_FUNCTION_ARGS)
 
 /*
  * The following operators support character-by-character comparison
- * of text data types, to allow building indexes suitable for LIKE
- * clauses.
+ * of text datums, to allow building indexes suitable for LIKE clauses.
+ * Note that the regular texteq/textne comparison operators are assumed
+ * to be compatible with these!
  */
 
 static int
 internal_text_pattern_compare(text *arg1, text *arg2)
 {
 	int			result;
+	int			len1,
+				len2;
+
+	len1 = VARSIZE_ANY_EXHDR(arg1);
+	len2 = VARSIZE_ANY_EXHDR(arg2);
 
-	result = memcmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2),
-					Min(VARSIZE_ANY_EXHDR(arg1), VARSIZE_ANY_EXHDR(arg2)));
+	result = strncmp(VARDATA_ANY(arg1), VARDATA_ANY(arg2), Min(len1, len2));
 	if (result != 0)
 		return result;
-	else if (VARSIZE_ANY_EXHDR(arg1) < VARSIZE_ANY_EXHDR(arg2))
+	else if (len1 < len2)
 		return -1;
-	else if (VARSIZE_ANY_EXHDR(arg1) > VARSIZE_ANY_EXHDR(arg2))
+	else if (len1 > len2)
 		return 1;
 	else
 		return 0;
@@ -1363,25 +1368,6 @@ text_pattern_le(PG_FUNCTION_ARGS)
 }
 
 
-Datum
-text_pattern_eq(PG_FUNCTION_ARGS)
-{
-	text	   *arg1 = PG_GETARG_TEXT_PP(0);
-	text	   *arg2 = PG_GETARG_TEXT_PP(1);
-	int			result;
-
-	if (VARSIZE_ANY_EXHDR(arg1) != VARSIZE_ANY_EXHDR(arg2))
-		result = 1;
-	else
-		result = internal_text_pattern_compare(arg1, arg2);
-
-	PG_FREE_IF_COPY(arg1, 0);
-	PG_FREE_IF_COPY(arg2, 1);
-
-	PG_RETURN_BOOL(result == 0);
-}
-
-
 Datum
 text_pattern_ge(PG_FUNCTION_ARGS)
 {
@@ -1414,25 +1400,6 @@ text_pattern_gt(PG_FUNCTION_ARGS)
 }
 
 
-Datum
-text_pattern_ne(PG_FUNCTION_ARGS)
-{
-	text	   *arg1 = PG_GETARG_TEXT_PP(0);
-	text	   *arg2 = PG_GETARG_TEXT_PP(1);
-	int			result;
-
-	if (VARSIZE_ANY_EXHDR(arg1) != VARSIZE_ANY_EXHDR(arg2))
-		result = 1;
-	else
-		result = internal_text_pattern_compare(arg1, arg2);
-
-	PG_FREE_IF_COPY(arg1, 0);
-	PG_FREE_IF_COPY(arg2, 1);
-
-	PG_RETURN_BOOL(result != 0);
-}
-
-
 Datum
 bttext_pattern_cmp(PG_FUNCTION_ARGS)
 {

src/include/catalog/catversion.h

@@ -37,7 +37,7 @@
  * Portions Copyright (c) 1996-2008, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California
  *
- * $PostgreSQL: pgsql/src/include/catalog/catversion.h,v 1.461 2008/05/16 23:36:05 tgl Exp $
+ * $PostgreSQL: pgsql/src/include/catalog/catversion.h,v 1.462 2008/05/27 00:13:09 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -53,6 +53,6 @@
  */
 
 /*							yyyymmddN */
-#define CATALOG_VERSION_NO	200805162
+#define CATALOG_VERSION_NO	200805261
 
 #endif