Skip to content

Commit 3187d6d

Browse files
feodorfeodor
committed
Introduce parse_ident()
SQL-layer function to split qualified identifier into array parts. Author: Pavel Stehule with minor editorization by me and Jim Nasby
1 parent 992b5ba commit 3187d6d

File tree

10 files changed

+375
-2
lines changed

10 files changed

+375
-2
lines changed

‎doc/src/sgml/func.sgml

Lines changed: 26 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1818,6 +1818,32 @@
18181818
<entry><literal>900150983cd24fb0 d6963f7d28e17f72</literal></entry>
18191819
</row>
18201820

1821+
<row>
1822+
<entry>
1823+
<indexterm>
1824+
<primary>parse_ident</primary>
1825+
</indexterm>
1826+
<literal><function>parse_ident(<parameter>str</parameter> <type>text</type>,
1827+
[ <parameter>strictmode</parameter> <type>boolean</type> DEFAULT true ] )</function></literal>
1828+
</entry>
1829+
<entry><type>text[]</type></entry>
1830+
<entry>Split <parameter>qualified identifier</parameter> into array
1831+
<parameter>parts</parameter>. When <parameter>strictmode</parameter> is
1832+
false, extra characters after the identifier are ignored. This is useful
1833+
for parsing identifiers for objects like functions and arrays that may
1834+
have trailing characters. By default, extra characters after the last
1835+
identifier are considered an error, but if second parameter is false,
1836+
then chararacters after last identifier are ignored. Note that this
1837+
function does not truncate quoted identifiers. If you care about that
1838+
you should cast the result of this function to name[]. A non-printable
1839+
chararacters (like 0 to 31) are displayed as hexadecimal codes always,
1840+
what can be different from PostgreSQL internal SQL identifiers
1841+
processing, when the original escaped value is displayed.
1842+
</entry>
1843+
<entry><literal>parse_ident('"SomeSchema".someTable')</literal></entry>
1844+
<entry><literal>"SomeSchema,sometable"</literal></entry>
1845+
</row>
1846+
18211847
<row>
18221848
<entry>
18231849
<indexterm>

‎src/backend/catalog/system_views.sql

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -990,3 +990,10 @@ RETURNS jsonb
990990
LANGUAGE INTERNAL
991991
STRICT IMMUTABLE
992992
AS 'jsonb_set';
993+
994+
CREATE OR REPLACE FUNCTION
995+
parse_ident(str text, strict boolean DEFAULT true)
996+
RETURNS text[]
997+
LANGUAGE INTERNAL
998+
STRICT IMMUTABLE
999+
AS 'parse_ident';

‎src/backend/parser/scansup.c

Lines changed: 11 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -129,6 +129,15 @@ scanstr(const char *s)
129129
*/
130130
char *
131131
downcase_truncate_identifier(const char *ident, int len, bool warn)
132+
{
133+
return downcase_identifier(ident, len, warn, true);
134+
}
135+
136+
/*
137+
* a workhorse for downcase_truncate_identifier
138+
*/
139+
char *
140+
downcase_identifier(const char *ident, int len, bool warn, bool truncate)
132141
{
133142
char *result;
134143
int i;
@@ -158,12 +167,13 @@ downcase_truncate_identifier(const char *ident, int len, bool warn)
158167
}
159168
result[i] = '\0';
160169

161-
if (i >= NAMEDATALEN)
170+
if (i >= NAMEDATALEN && truncate)
162171
truncate_identifier(result, i, warn);
163172

164173
return result;
165174
}
166175

176+
167177
/*
168178
* truncate_identifier() --- truncate an identifier to NAMEDATALEN-1 bytes.
169179
*

‎src/backend/utils/adt/misc.c

Lines changed: 224 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -27,6 +27,7 @@
2727
#include "commands/dbcommands.h"
2828
#include "funcapi.h"
2929
#include "miscadmin.h"
30+
#include "parser/scansup.h"
3031
#include "parser/keywords.h"
3132
#include "postmaster/syslogger.h"
3233
#include "rewrite/rewriteHandler.h"
@@ -719,3 +720,226 @@ pg_column_is_updatable(PG_FUNCTION_ARGS)
719720

720721
PG_RETURN_BOOL((events & REQ_EVENTS) == REQ_EVENTS);
721722
}
723+
724+
725+
/*
726+
* This simple parser utility are compatible with lexer implementation,
727+
* used only in parse_ident function
728+
*/
729+
static bool
730+
is_ident_start(unsigned char c)
731+
{
732+
if (c == '_')
733+
return true;
734+
if ((c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z'))
735+
return true;
736+
737+
if (c >= 0200 && c <= 0377)
738+
return true;
739+
740+
return false;
741+
}
742+
743+
static bool
744+
is_ident_cont(unsigned char c)
745+
{
746+
if (c >= '0' && c <= '9')
747+
return true;
748+
749+
return is_ident_start(c);
750+
}
751+
752+
/*
753+
* Sanitize SQL string for using in error message.
754+
*/
755+
static char *
756+
sanitize_text(text *t)
757+
{
758+
int len = VARSIZE_ANY_EXHDR(t);
759+
const char *p = VARDATA_ANY(t);
760+
StringInfo dstr;
761+
762+
dstr = makeStringInfo();
763+
764+
appendStringInfoChar(dstr, '"');
765+
766+
while (len--)
767+
{
768+
switch (*p)
769+
{
770+
case '\b':
771+
appendStringInfoString(dstr, "\\b");
772+
break;
773+
case '\f':
774+
appendStringInfoString(dstr, "\\f");
775+
break;
776+
case '\n':
777+
appendStringInfoString(dstr, "\\n");
778+
break;
779+
case '\r':
780+
appendStringInfoString(dstr, "\\r");
781+
break;
782+
case '\t':
783+
appendStringInfoString(dstr, "\\t");
784+
break;
785+
case '\'':
786+
appendStringInfoString(dstr, "''");
787+
break;
788+
case '\\':
789+
appendStringInfoString(dstr, "\\\\");
790+
break;
791+
default:
792+
if ((unsigned char) *p < ' ')
793+
appendStringInfo(dstr, "\\u%04x", (int) *p);
794+
else
795+
appendStringInfoCharMacro(dstr, *p);
796+
break;
797+
}
798+
p++;
799+
}
800+
801+
appendStringInfoChar(dstr, '"');
802+
803+
return dstr->data;
804+
}
805+
806+
/*
807+
* parse_ident - parse SQL composed identifier to separate identifiers.
808+
* When strict mode is active (second parameter), then any chars after
809+
* last identifiers are disallowed.
810+
*/
811+
Datum
812+
parse_ident(PG_FUNCTION_ARGS)
813+
{
814+
text *qualname;
815+
char *qualname_str;
816+
bool strict;
817+
char *nextp;
818+
bool after_dot = false;
819+
ArrayBuildState *astate = NULL;
820+
821+
qualname = PG_GETARG_TEXT_PP(0);
822+
qualname_str = text_to_cstring(qualname);
823+
strict = PG_GETARG_BOOL(1);
824+
825+
nextp = qualname_str;
826+
827+
/* skip leading whitespace */
828+
while (isspace((unsigned char) *nextp))
829+
nextp++;
830+
831+
for (;;)
832+
{
833+
char *curname;
834+
char *endp;
835+
bool missing_ident;
836+
837+
missing_ident = true;
838+
839+
if (*nextp == '\"')
840+
{
841+
curname = nextp + 1;
842+
for (;;)
843+
{
844+
endp = strchr(nextp + 1, '\"');
845+
if (endp == NULL)
846+
ereport(ERROR,
847+
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
848+
errmsg("unclosed double quotes"),
849+
errdetail("string %s is not valid identifier",
850+
sanitize_text(qualname))));
851+
if (endp[1] != '\"')
852+
break;
853+
memmove(endp, endp + 1, strlen(endp));
854+
nextp = endp;
855+
}
856+
nextp = endp + 1;
857+
*endp = '\0';
858+
859+
/* Show complete input string in this case. */
860+
if (endp - curname == 0)
861+
ereport(ERROR,
862+
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
863+
errmsg("identifier should not be empty: %s",
864+
sanitize_text(qualname))));
865+
866+
astate = accumArrayResult(astate, CStringGetTextDatum(curname),
867+
false, TEXTOID, CurrentMemoryContext);
868+
missing_ident = false;
869+
}
870+
else
871+
{
872+
if (is_ident_start((unsigned char) *nextp))
873+
{
874+
char *downname;
875+
int len;
876+
text *part;
877+
878+
curname = nextp++;
879+
while (is_ident_cont((unsigned char) *nextp))
880+
nextp++;
881+
882+
len = nextp - curname;
883+
884+
/*
885+
* Unlike name, we don't implicitly truncate identifiers. This
886+
* is useful for allowing the user to check for specific parts
887+
* of the identifier being too long. It's easy enough for the
888+
* user to get the truncated names by casting our output to
889+
* name[].
890+
*/
891+
downname = downcase_identifier(curname, len, false, false);
892+
part = cstring_to_text_with_len(downname, len);
893+
astate = accumArrayResult(astate, PointerGetDatum(part), false,
894+
TEXTOID, CurrentMemoryContext);
895+
missing_ident = false;
896+
}
897+
}
898+
899+
if (missing_ident)
900+
{
901+
/* Different error messages based on where we failed. */
902+
if (*nextp == '.')
903+
ereport(ERROR,
904+
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
905+
errmsg("missing valid identifier before \".\" symbol: %s",
906+
sanitize_text(qualname))));
907+
else if (after_dot)
908+
ereport(ERROR,
909+
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
910+
errmsg("missing valid identifier after \".\" symbol: %s",
911+
sanitize_text(qualname))));
912+
else
913+
ereport(ERROR,
914+
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
915+
errmsg("missing valid identifier: %s",
916+
sanitize_text(qualname))));
917+
}
918+
919+
while (isspace((unsigned char) *nextp))
920+
nextp++;
921+
922+
if (*nextp == '.')
923+
{
924+
after_dot = true;
925+
nextp++;
926+
while (isspace((unsigned char) *nextp))
927+
nextp++;
928+
}
929+
else if (*nextp == '\0')
930+
{
931+
break;
932+
}
933+
else
934+
{
935+
if (strict)
936+
ereport(ERROR,
937+
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
938+
errmsg("identifier contains disallowed characters: %s",
939+
sanitize_text(qualname))));
940+
break;
941+
}
942+
}
943+
944+
PG_RETURN_DATUM(makeArrayResult(astate, CurrentMemoryContext));
945+
}

‎src/include/catalog/catversion.h

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -53,6 +53,6 @@
5353
*/
5454

5555
/* yyyymmddN */
56-
#define CATALOG_VERSION_NO 201603151
56+
#define CATALOG_VERSION_NO 201603181
5757

5858
#endif

‎src/include/catalog/pg_proc.h

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -3463,6 +3463,9 @@ DESCR("I/O");
34633463
DATA(insert OID = 4086 ( to_regnamespace PGNSP PGUID 12 1 0 0 0 f f f f t f s s 1 0 4089 "25" _null_ _null_ _null_ _null_ _null_ to_regnamespace _null_ _null_ _null_ ));
34643464
DESCR("convert namespace name to regnamespace");
34653465

3466+
DATA(insert OID = 1268 ( parse_ident PGNSP PGUID 12 1 0 0 0 f f f f t f i s 2 0 1009 "25 16" _null_ _null_ "{str,strict}" _null_ _null_ parse_ident _null_ _null_ _null_ ));
3467+
DESCR("parse qualified identifier to array of identifiers");
3468+
34663469
DATA(insert OID = 2246 ( fmgr_internal_validator PGNSP PGUID 12 1 0 0 0 f f f f t f s s 1 0 2278 "26" _null_ _null_ _null_ _null_ _null_ fmgr_internal_validator _null_ _null_ _null_ ));
34673470
DESCR("(internal)");
34683471
DATA(insert OID = 2247 ( fmgr_c_validator PGNSP PGUID 12 1 0 0 0 f f f f t f s s 1 0 2278 "26" _null_ _null_ _null_ _null_ _null_ fmgr_c_validator _null_ _null_ _null_ ));

‎src/include/parser/scansup.h

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -20,6 +20,9 @@ extern char *scanstr(const char *s);
2020
extern char *downcase_truncate_identifier(const char *ident, int len,
2121
bool warn);
2222

23+
extern char *downcase_identifier(const char *ident, int len,
24+
bool warn, bool truncate);
25+
2326
extern void truncate_identifier(char *ident, int len, bool warn);
2427

2528
extern bool scanner_isspace(char ch);

‎src/include/utils/builtins.h

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -510,6 +510,7 @@ extern Datum pg_typeof(PG_FUNCTION_ARGS);
510510
extern Datum pg_collation_for(PG_FUNCTION_ARGS);
511511
extern Datum pg_relation_is_updatable(PG_FUNCTION_ARGS);
512512
extern Datum pg_column_is_updatable(PG_FUNCTION_ARGS);
513+
extern Datum parse_ident(PG_FUNCTION_ARGS);
513514

514515
/* oid.c */
515516
extern Datum oidin(PG_FUNCTION_ARGS);

0 commit comments

Comments
 (0)