Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commitd8f3752

Browse files
committed
Generate double-sided LIKE indexquals that work even in weird locales,
by continuing to increment the rightmost character until we get a stringthat is demonstrably greater than the pattern prefix.
1 parent5f68d5c commitd8f3752

File tree

1 file changed

+163
-45
lines changed

1 file changed

+163
-45
lines changed

‎src/backend/optimizer/path/indxpath.c‎

Lines changed: 163 additions & 45 deletions
Original file line numberDiff line numberDiff line change
@@ -8,7 +8,7 @@
88
*
99
*
1010
* IDENTIFICATION
11-
* $Header: /cvsroot/pgsql/src/backend/optimizer/path/indxpath.c,v 1.74 1999/12/3103:41:03 tgl Exp $
11+
* $Header: /cvsroot/pgsql/src/backend/optimizer/path/indxpath.c,v 1.75 1999/12/3105:38:25 tgl Exp $
1212
*
1313
*-------------------------------------------------------------------------
1414
*/
@@ -23,6 +23,7 @@
2323
#include"catalog/pg_amop.h"
2424
#include"catalog/pg_operator.h"
2525
#include"executor/executor.h"
26+
#include"mb/pg_wchar.h"
2627
#include"nodes/makefuncs.h"
2728
#include"nodes/nodeFuncs.h"
2829
#include"optimizer/clauses.h"
@@ -92,7 +93,12 @@ static Prefix_Status regex_fixed_prefix(char *patt, bool case_insensitive,
9293
char**prefix);
9394
staticList*prefix_quals(Var*leftop,Oidexpr_op,
9495
char*prefix,Prefix_Statuspstatus);
96+
staticchar*make_greater_string(constchar*str,Oiddatatype);
9597
staticOidfind_operator(constchar*opname,Oiddatatype);
98+
staticDatumstring_to_datum(constchar*str,Oiddatatype);
99+
staticConst*string_to_const(constchar*str,Oiddatatype);
100+
staticboolstring_lessthan(constchar*str1,constchar*str2,
101+
Oiddatatype);
96102

97103

98104
/*
@@ -1653,31 +1659,31 @@ match_special_index_operator(Expr *clause, Oid opclass, Oid relam,
16531659
caseOID_TEXT_REGEXEQ_OP:
16541660
caseOID_TEXT_ICREGEXEQ_OP:
16551661
if (!op_class(find_operator(">=",TEXTOID),opclass,relam)||
1656-
!op_class(find_operator("<=",TEXTOID),opclass,relam))
1662+
!op_class(find_operator("<",TEXTOID),opclass,relam))
16571663
isIndexable= false;
16581664
break;
16591665

16601666
caseOID_BPCHAR_LIKE_OP:
16611667
caseOID_BPCHAR_REGEXEQ_OP:
16621668
caseOID_BPCHAR_ICREGEXEQ_OP:
16631669
if (!op_class(find_operator(">=",BPCHAROID),opclass,relam)||
1664-
!op_class(find_operator("<=",BPCHAROID),opclass,relam))
1670+
!op_class(find_operator("<",BPCHAROID),opclass,relam))
16651671
isIndexable= false;
16661672
break;
16671673

16681674
caseOID_VARCHAR_LIKE_OP:
16691675
caseOID_VARCHAR_REGEXEQ_OP:
16701676
caseOID_VARCHAR_ICREGEXEQ_OP:
16711677
if (!op_class(find_operator(">=",VARCHAROID),opclass,relam)||
1672-
!op_class(find_operator("<=",VARCHAROID),opclass,relam))
1678+
!op_class(find_operator("<",VARCHAROID),opclass,relam))
16731679
isIndexable= false;
16741680
break;
16751681

16761682
caseOID_NAME_LIKE_OP:
16771683
caseOID_NAME_REGEXEQ_OP:
16781684
caseOID_NAME_ICREGEXEQ_OP:
16791685
if (!op_class(find_operator(">=",NAMEOID),opclass,relam)||
1680-
!op_class(find_operator("<=",NAMEOID),opclass,relam))
1686+
!op_class(find_operator("<",NAMEOID),opclass,relam))
16811687
isIndexable= false;
16821688
break;
16831689
}
@@ -1774,7 +1780,7 @@ expand_indexqual_conditions(List *indexquals)
17741780

17751781
/*
17761782
* Extract the fixed prefix, if any, for a LIKE pattern.
1777-
* *prefix is set to a palloc'd prefix string with 1 spare byte,
1783+
* *prefix is set to a palloc'd prefix string,
17781784
* or to NULL if no fixed prefix exists for the pattern.
17791785
* The return value distinguishes no fixed prefix, a partial prefix,
17801786
* or an exact-match-only pattern.
@@ -1786,7 +1792,7 @@ like_fixed_prefix(char *patt, char **prefix)
17861792
intpos,
17871793
match_pos;
17881794

1789-
*prefix=match=palloc(strlen(patt)+2);
1795+
*prefix=match=palloc(strlen(patt)+1);
17901796
match_pos=0;
17911797

17921798
for (pos=0;patt[pos];pos++)
@@ -1823,7 +1829,7 @@ like_fixed_prefix(char *patt, char **prefix)
18231829

18241830
/*
18251831
* Extract the fixed prefix, if any, for a regex pattern.
1826-
* *prefix is set to a palloc'd prefix string with 1 spare byte,
1832+
* *prefix is set to a palloc'd prefix string,
18271833
* or to NULL if no fixed prefix exists for the pattern.
18281834
* The return value distinguishes no fixed prefix, a partial prefix,
18291835
* or an exact-match-only pattern.
@@ -1858,7 +1864,7 @@ regex_fixed_prefix(char *patt, bool case_insensitive,
18581864
}
18591865

18601866
/* OK, allocate space for pattern */
1861-
*prefix=match=palloc(strlen(patt)+2);
1867+
*prefix=match=palloc(strlen(patt)+1);
18621868
match_pos=0;
18631869

18641870
/* note start at pos 1 to skip leading ^ */
@@ -1906,11 +1912,10 @@ prefix_quals(Var *leftop, Oid expr_op,
19061912
List*result;
19071913
Oiddatatype;
19081914
Oidoproid;
1909-
void*conval;
19101915
Const*con;
19111916
Oper*op;
19121917
Expr*expr;
1913-
intprefixlen;
1918+
char*greaterstr;
19141919

19151920
Assert(pstatus!=Prefix_None);
19161921

@@ -1953,14 +1958,7 @@ prefix_quals(Var *leftop, Oid expr_op,
19531958
oproid=find_operator("=",datatype);
19541959
if (oproid==InvalidOid)
19551960
elog(ERROR,"prefix_quals: no = operator for type %u",datatype);
1956-
/* Note: we cheat a little by assuming that textin() will do for
1957-
* bpchar and varchar constants too...
1958-
*/
1959-
conval= (datatype==NAMEOID) ?
1960-
(void*)namein(prefix) : (void*)textin(prefix);
1961-
con=makeConst(datatype, ((datatype==NAMEOID) ?NAMEDATALEN :-1),
1962-
PointerGetDatum(conval),
1963-
false, false, false, false);
1961+
con=string_to_const(prefix,datatype);
19641962
op=makeOper(oproid,InvalidOid,BOOLOID,0,NULL);
19651963
expr=make_opclause(op,leftop, (Var*)con);
19661964
result=lcons(expr,NIL);
@@ -1975,43 +1973,92 @@ prefix_quals(Var *leftop, Oid expr_op,
19751973
oproid=find_operator(">=",datatype);
19761974
if (oproid==InvalidOid)
19771975
elog(ERROR,"prefix_quals: no >= operator for type %u",datatype);
1978-
conval= (datatype==NAMEOID) ?
1979-
(void*)namein(prefix) : (void*)textin(prefix);
1980-
con=makeConst(datatype, ((datatype==NAMEOID) ?NAMEDATALEN :-1),
1981-
PointerGetDatum(conval),
1982-
false, false, false, false);
1976+
con=string_to_const(prefix,datatype);
19831977
op=makeOper(oproid,InvalidOid,BOOLOID,0,NULL);
19841978
expr=make_opclause(op,leftop, (Var*)con);
19851979
result=lcons(expr,NIL);
19861980

19871981
/*
1988-
* In ASCII locale we say "x <= prefix\377". This does not
1989-
* work for non-ASCII collation orders, and it's not really
1990-
* right even for ASCII. FIX ME!
1991-
* Note we assume the passed prefix string is workspace with
1992-
* an extra byte, as created by the xxx_fixed_prefix routines above.
1982+
* If we can create a string larger than the prefix, say "x < greaterstr".
19931983
*/
1994-
#ifndefUSE_LOCALE
1995-
prefixlen=strlen(prefix);
1996-
prefix[prefixlen]='\377';
1997-
prefix[prefixlen+1]='\0';
1984+
greaterstr=make_greater_string(prefix,datatype);
1985+
if (greaterstr)
1986+
{
1987+
oproid=find_operator("<",datatype);
1988+
if (oproid==InvalidOid)
1989+
elog(ERROR,"prefix_quals: no < operator for type %u",datatype);
1990+
con=string_to_const(greaterstr,datatype);
1991+
op=makeOper(oproid,InvalidOid,BOOLOID,0,NULL);
1992+
expr=make_opclause(op,leftop, (Var*)con);
1993+
result=lappend(result,expr);
1994+
pfree(greaterstr);
1995+
}
19981996

1999-
oproid=find_operator("<=",datatype);
2000-
if (oproid==InvalidOid)
2001-
elog(ERROR,"prefix_quals: no <= operator for type %u",datatype);
2002-
conval= (datatype==NAMEOID) ?
2003-
(void*)namein(prefix) : (void*)textin(prefix);
2004-
con=makeConst(datatype, ((datatype==NAMEOID) ?NAMEDATALEN :-1),
2005-
PointerGetDatum(conval),
2006-
false, false, false, false);
2007-
op=makeOper(oproid,InvalidOid,BOOLOID,0,NULL);
2008-
expr=make_opclause(op,leftop, (Var*)con);
2009-
result=lappend(result,expr);
1997+
returnresult;
1998+
}
1999+
2000+
/*
2001+
* Try to generate a string greater than the given string or any string it is
2002+
* a prefix of. If successful, return a palloc'd string; else return NULL.
2003+
*
2004+
* To work correctly in non-ASCII locales with weird collation orders,
2005+
* we cannot simply increment "foo" to "fop" --- we have to check whether
2006+
* we actually produced a string greater than the given one. If not,
2007+
* increment the righthand byte again and repeat. If we max out the righthand
2008+
* byte, truncate off the last character and start incrementing the next.
2009+
* For example, if "z" were the last character in the sort order, then we
2010+
* could produce "foo" as a string greater than "fonz".
2011+
*
2012+
* This could be rather slow in the worst case, but in most cases we won't
2013+
* have to try more than one or two strings before succeeding.
2014+
*
2015+
* XXX in a sufficiently weird locale, this might produce incorrect results?
2016+
* For example, in German I believe "ss" is treated specially --- if we are
2017+
* given "foos" and return "foot", will this actually be greater than "fooss"?
2018+
*/
2019+
staticchar*
2020+
make_greater_string(constchar*str,Oiddatatype)
2021+
{
2022+
char*workstr;
2023+
intlen;
2024+
2025+
/* Make a modifiable copy, which will be our return value if successful */
2026+
workstr=pstrdup((char*)str);
2027+
2028+
while ((len=strlen(workstr))>0)
2029+
{
2030+
unsignedchar*lastchar= (unsignedchar*) (workstr+len-1);
2031+
2032+
/*
2033+
* Try to generate a larger string by incrementing the last byte.
2034+
*/
2035+
while (*lastchar< (unsignedchar)255)
2036+
{
2037+
(*lastchar)++;
2038+
if (string_lessthan(str,workstr,datatype))
2039+
returnworkstr;/* Success! */
2040+
}
2041+
/*
2042+
* Truncate off the last character, which might be more than 1 byte
2043+
* in MULTIBYTE case.
2044+
*/
2045+
#ifdefMULTIBYTE
2046+
len=pg_mbcliplen((constunsignedchar*)workstr,len,len-1);
2047+
workstr[len]='\0';
2048+
#else
2049+
*lastchar='\0';
20102050
#endif
2051+
}
20112052

2012-
returnresult;
2053+
/* Failed... */
2054+
pfree(workstr);
2055+
returnNULL;
20132056
}
20142057

2058+
/*
2059+
* Handy subroutines for match_special_index_operator() and friends.
2060+
*/
2061+
20152062
/* See if there is a binary op of the given name for the given datatype */
20162063
staticOid
20172064
find_operator(constchar*opname,Oiddatatype)
@@ -2027,3 +2074,74 @@ find_operator(const char * opname, Oid datatype)
20272074
returnInvalidOid;
20282075
returnoptup->t_data->t_oid;
20292076
}
2077+
2078+
/*
2079+
* Generate a Datum of the appropriate type from a C string.
2080+
* Note that all of the supported types are pass-by-ref, so the
2081+
* returned value should be pfree'd if no longer needed.
2082+
*/
2083+
staticDatum
2084+
string_to_datum(constchar*str,Oiddatatype)
2085+
{
2086+
/* We cheat a little by assuming that textin() will do for
2087+
* bpchar and varchar constants too...
2088+
*/
2089+
if (datatype==NAMEOID)
2090+
returnPointerGetDatum(namein((char*)str));
2091+
else
2092+
returnPointerGetDatum(textin((char*)str));
2093+
}
2094+
2095+
/*
2096+
* Generate a Const node of the appropriate type from a C string.
2097+
*/
2098+
staticConst*
2099+
string_to_const(constchar*str,Oiddatatype)
2100+
{
2101+
Datumconval=string_to_datum(str,datatype);
2102+
2103+
returnmakeConst(datatype, ((datatype==NAMEOID) ?NAMEDATALEN :-1),
2104+
conval, false, false, false, false);
2105+
}
2106+
2107+
/*
2108+
* Test whether two strings are "<" according to the rules of the given
2109+
* datatype. We do this the hard way, ie, actually calling the type's
2110+
* "<" operator function, to ensure we get the right result...
2111+
*/
2112+
staticbool
2113+
string_lessthan(constchar*str1,constchar*str2,Oiddatatype)
2114+
{
2115+
Datumdatum1=string_to_datum(str1,datatype);
2116+
Datumdatum2=string_to_datum(str2,datatype);
2117+
boolresult;
2118+
2119+
switch (datatype)
2120+
{
2121+
caseTEXTOID:
2122+
result=text_lt((text*)datum1, (text*)datum2);
2123+
break;
2124+
2125+
caseBPCHAROID:
2126+
result=bpcharlt((char*)datum1, (char*)datum2);
2127+
break;
2128+
2129+
caseVARCHAROID:
2130+
result=varcharlt((char*)datum1, (char*)datum2);
2131+
break;
2132+
2133+
caseNAMEOID:
2134+
result=namelt((NameData*)datum1, (NameData*)datum2);
2135+
break;
2136+
2137+
default:
2138+
elog(ERROR,"string_lessthan: unexpected datatype %u",datatype);
2139+
result= false;
2140+
break;
2141+
}
2142+
2143+
pfree(DatumGetPointer(datum1));
2144+
pfree(DatumGetPointer(datum2));
2145+
2146+
returnresult;
2147+
}

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp