summaryrefslogtreecommitdiffstats
path: root/contrib/dict_int
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--contrib/dict_int/.gitignore4
-rw-r--r--contrib/dict_int/Makefile23
-rw-r--r--contrib/dict_int/dict_int--1.0.sql25
-rw-r--r--contrib/dict_int/dict_int.c116
-rw-r--r--contrib/dict_int/dict_int.control6
-rw-r--r--contrib/dict_int/expected/dict_int.out356
-rw-r--r--contrib/dict_int/sql/dict_int.sql69
7 files changed, 599 insertions, 0 deletions
diff --git a/contrib/dict_int/.gitignore b/contrib/dict_int/.gitignore
new file mode 100644
index 0000000..5dcb3ff
--- /dev/null
+++ b/contrib/dict_int/.gitignore
@@ -0,0 +1,4 @@
+# Generated subdirectories
+/log/
+/results/
+/tmp_check/
diff --git a/contrib/dict_int/Makefile b/contrib/dict_int/Makefile
new file mode 100644
index 0000000..2ff1bdb
--- /dev/null
+++ b/contrib/dict_int/Makefile
@@ -0,0 +1,23 @@
+# contrib/dict_int/Makefile
+
+MODULE_big = dict_int
+OBJS = \
+ $(WIN32RES) \
+ dict_int.o
+
+EXTENSION = dict_int
+DATA = dict_int--1.0.sql
+PGFILEDESC = "dict_int - add-on dictionary template for full-text search"
+
+REGRESS = dict_int
+
+ifdef USE_PGXS
+PG_CONFIG = pg_config
+PGXS := $(shell $(PG_CONFIG) --pgxs)
+include $(PGXS)
+else
+subdir = contrib/dict_int
+top_builddir = ../..
+include $(top_builddir)/src/Makefile.global
+include $(top_srcdir)/contrib/contrib-global.mk
+endif
diff --git a/contrib/dict_int/dict_int--1.0.sql b/contrib/dict_int/dict_int--1.0.sql
new file mode 100644
index 0000000..acb1461
--- /dev/null
+++ b/contrib/dict_int/dict_int--1.0.sql
@@ -0,0 +1,25 @@
+/* contrib/dict_int/dict_int--1.0.sql */
+
+-- complain if script is sourced in psql, rather than via CREATE EXTENSION
+\echo Use "CREATE EXTENSION dict_int" to load this file. \quit
+
+CREATE FUNCTION dintdict_init(internal)
+ RETURNS internal
+ AS 'MODULE_PATHNAME'
+ LANGUAGE C STRICT;
+
+CREATE FUNCTION dintdict_lexize(internal, internal, internal, internal)
+ RETURNS internal
+ AS 'MODULE_PATHNAME'
+ LANGUAGE C STRICT;
+
+CREATE TEXT SEARCH TEMPLATE intdict_template (
+ LEXIZE = dintdict_lexize,
+ INIT = dintdict_init
+);
+
+CREATE TEXT SEARCH DICTIONARY intdict (
+ TEMPLATE = intdict_template
+);
+
+COMMENT ON TEXT SEARCH DICTIONARY intdict IS 'dictionary for integers';
diff --git a/contrib/dict_int/dict_int.c b/contrib/dict_int/dict_int.c
new file mode 100644
index 0000000..006bacb
--- /dev/null
+++ b/contrib/dict_int/dict_int.c
@@ -0,0 +1,116 @@
+/*-------------------------------------------------------------------------
+ *
+ * dict_int.c
+ * Text search dictionary for integers
+ *
+ * Copyright (c) 2007-2022, PostgreSQL Global Development Group
+ *
+ * IDENTIFICATION
+ * contrib/dict_int/dict_int.c
+ *
+ *-------------------------------------------------------------------------
+ */
+#include "postgres.h"
+
+#include "commands/defrem.h"
+#include "tsearch/ts_public.h"
+
+PG_MODULE_MAGIC;
+
+typedef struct
+{
+ int maxlen;
+ bool rejectlong;
+ bool absval;
+} DictInt;
+
+
+PG_FUNCTION_INFO_V1(dintdict_init);
+PG_FUNCTION_INFO_V1(dintdict_lexize);
+
+Datum
+dintdict_init(PG_FUNCTION_ARGS)
+{
+ List *dictoptions = (List *) PG_GETARG_POINTER(0);
+ DictInt *d;
+ ListCell *l;
+
+ d = (DictInt *) palloc0(sizeof(DictInt));
+ d->maxlen = 6;
+ d->rejectlong = false;
+ d->absval = false;
+
+ foreach(l, dictoptions)
+ {
+ DefElem *defel = (DefElem *) lfirst(l);
+
+ if (strcmp(defel->defname, "maxlen") == 0)
+ {
+ d->maxlen = atoi(defGetString(defel));
+
+ if (d->maxlen < 1)
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("maxlen value has to be >= 1")));
+ }
+ else if (strcmp(defel->defname, "rejectlong") == 0)
+ {
+ d->rejectlong = defGetBoolean(defel);
+ }
+ else if (strcmp(defel->defname, "absval") == 0)
+ {
+ d->absval = defGetBoolean(defel);
+ }
+ else
+ {
+ ereport(ERROR,
+ (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ errmsg("unrecognized intdict parameter: \"%s\"",
+ defel->defname)));
+ }
+ }
+
+ PG_RETURN_POINTER(d);
+}
+
+Datum
+dintdict_lexize(PG_FUNCTION_ARGS)
+{
+ DictInt *d = (DictInt *) PG_GETARG_POINTER(0);
+ char *in = (char *) PG_GETARG_POINTER(1);
+ int len = PG_GETARG_INT32(2);
+ char *txt;
+ TSLexeme *res = palloc0(sizeof(TSLexeme) * 2);
+
+ res[1].lexeme = NULL;
+
+ if (d->absval && (in[0] == '+' || in[0] == '-'))
+ {
+ len--;
+ txt = pnstrdup(in + 1, len);
+ }
+ else
+ txt = pnstrdup(in, len);
+
+ if (len > d->maxlen)
+ {
+ if (d->rejectlong)
+ {
+ /* reject by returning void array */
+ pfree(txt);
+ res[0].lexeme = NULL;
+ }
+ else
+ {
+ /* trim integer */
+ txt[d->maxlen] = '\0';
+ res[0].lexeme = txt;
+ }
+ }
+ else
+ {
+ res[0].lexeme = txt;
+ }
+
+ PG_RETURN_POINTER(res);
+}
diff --git a/contrib/dict_int/dict_int.control b/contrib/dict_int/dict_int.control
new file mode 100644
index 0000000..ec04cce
--- /dev/null
+++ b/contrib/dict_int/dict_int.control
@@ -0,0 +1,6 @@
+# dict_int extension
+comment = 'text search dictionary template for integers'
+default_version = '1.0'
+module_pathname = '$libdir/dict_int'
+relocatable = true
+trusted = true
diff --git a/contrib/dict_int/expected/dict_int.out b/contrib/dict_int/expected/dict_int.out
new file mode 100644
index 0000000..39906fc
--- /dev/null
+++ b/contrib/dict_int/expected/dict_int.out
@@ -0,0 +1,356 @@
+CREATE EXTENSION dict_int;
+--lexize
+select ts_lexize('intdict', '511673');
+ ts_lexize
+-----------
+ {511673}
+(1 row)
+
+select ts_lexize('intdict', '129');
+ ts_lexize
+-----------
+ {129}
+(1 row)
+
+select ts_lexize('intdict', '40865854');
+ ts_lexize
+-----------
+ {408658}
+(1 row)
+
+select ts_lexize('intdict', '952');
+ ts_lexize
+-----------
+ {952}
+(1 row)
+
+select ts_lexize('intdict', '654980341');
+ ts_lexize
+-----------
+ {654980}
+(1 row)
+
+select ts_lexize('intdict', '09810106');
+ ts_lexize
+-----------
+ {098101}
+(1 row)
+
+select ts_lexize('intdict', '14262713');
+ ts_lexize
+-----------
+ {142627}
+(1 row)
+
+select ts_lexize('intdict', '6532082986');
+ ts_lexize
+-----------
+ {653208}
+(1 row)
+
+select ts_lexize('intdict', '0150061');
+ ts_lexize
+-----------
+ {015006}
+(1 row)
+
+select ts_lexize('intdict', '7778');
+ ts_lexize
+-----------
+ {7778}
+(1 row)
+
+select ts_lexize('intdict', '9547');
+ ts_lexize
+-----------
+ {9547}
+(1 row)
+
+select ts_lexize('intdict', '753395478');
+ ts_lexize
+-----------
+ {753395}
+(1 row)
+
+select ts_lexize('intdict', '647652');
+ ts_lexize
+-----------
+ {647652}
+(1 row)
+
+select ts_lexize('intdict', '6988655574');
+ ts_lexize
+-----------
+ {698865}
+(1 row)
+
+select ts_lexize('intdict', '1279');
+ ts_lexize
+-----------
+ {1279}
+(1 row)
+
+select ts_lexize('intdict', '1266645909');
+ ts_lexize
+-----------
+ {126664}
+(1 row)
+
+select ts_lexize('intdict', '7594193969');
+ ts_lexize
+-----------
+ {759419}
+(1 row)
+
+select ts_lexize('intdict', '16928207');
+ ts_lexize
+-----------
+ {169282}
+(1 row)
+
+select ts_lexize('intdict', '196850350328');
+ ts_lexize
+-----------
+ {196850}
+(1 row)
+
+select ts_lexize('intdict', '22026985592');
+ ts_lexize
+-----------
+ {220269}
+(1 row)
+
+select ts_lexize('intdict', '2063765');
+ ts_lexize
+-----------
+ {206376}
+(1 row)
+
+select ts_lexize('intdict', '242387310');
+ ts_lexize
+-----------
+ {242387}
+(1 row)
+
+select ts_lexize('intdict', '93595');
+ ts_lexize
+-----------
+ {93595}
+(1 row)
+
+select ts_lexize('intdict', '9374');
+ ts_lexize
+-----------
+ {9374}
+(1 row)
+
+select ts_lexize('intdict', '996969');
+ ts_lexize
+-----------
+ {996969}
+(1 row)
+
+select ts_lexize('intdict', '353595982');
+ ts_lexize
+-----------
+ {353595}
+(1 row)
+
+select ts_lexize('intdict', '925860');
+ ts_lexize
+-----------
+ {925860}
+(1 row)
+
+select ts_lexize('intdict', '11848378337');
+ ts_lexize
+-----------
+ {118483}
+(1 row)
+
+select ts_lexize('intdict', '333');
+ ts_lexize
+-----------
+ {333}
+(1 row)
+
+select ts_lexize('intdict', '799287416765');
+ ts_lexize
+-----------
+ {799287}
+(1 row)
+
+select ts_lexize('intdict', '745939');
+ ts_lexize
+-----------
+ {745939}
+(1 row)
+
+select ts_lexize('intdict', '67601305734');
+ ts_lexize
+-----------
+ {676013}
+(1 row)
+
+select ts_lexize('intdict', '3361113');
+ ts_lexize
+-----------
+ {336111}
+(1 row)
+
+select ts_lexize('intdict', '9033778607');
+ ts_lexize
+-----------
+ {903377}
+(1 row)
+
+select ts_lexize('intdict', '7507648');
+ ts_lexize
+-----------
+ {750764}
+(1 row)
+
+select ts_lexize('intdict', '1166');
+ ts_lexize
+-----------
+ {1166}
+(1 row)
+
+select ts_lexize('intdict', '9360498');
+ ts_lexize
+-----------
+ {936049}
+(1 row)
+
+select ts_lexize('intdict', '917795');
+ ts_lexize
+-----------
+ {917795}
+(1 row)
+
+select ts_lexize('intdict', '9387894');
+ ts_lexize
+-----------
+ {938789}
+(1 row)
+
+select ts_lexize('intdict', '42764329');
+ ts_lexize
+-----------
+ {427643}
+(1 row)
+
+select ts_lexize('intdict', '564062');
+ ts_lexize
+-----------
+ {564062}
+(1 row)
+
+select ts_lexize('intdict', '5413377');
+ ts_lexize
+-----------
+ {541337}
+(1 row)
+
+select ts_lexize('intdict', '060965');
+ ts_lexize
+-----------
+ {060965}
+(1 row)
+
+select ts_lexize('intdict', '08273593');
+ ts_lexize
+-----------
+ {082735}
+(1 row)
+
+select ts_lexize('intdict', '593556010144');
+ ts_lexize
+-----------
+ {593556}
+(1 row)
+
+select ts_lexize('intdict', '17988843352');
+ ts_lexize
+-----------
+ {179888}
+(1 row)
+
+select ts_lexize('intdict', '252281774');
+ ts_lexize
+-----------
+ {252281}
+(1 row)
+
+select ts_lexize('intdict', '313425');
+ ts_lexize
+-----------
+ {313425}
+(1 row)
+
+select ts_lexize('intdict', '641439323669');
+ ts_lexize
+-----------
+ {641439}
+(1 row)
+
+select ts_lexize('intdict', '314532610153');
+ ts_lexize
+-----------
+ {314532}
+(1 row)
+
+ALTER TEXT SEARCH DICTIONARY intdict (MAXLEN = -214783648); -- fail
+ERROR: maxlen value has to be >= 1
+-- This ought to fail, perhaps, but historically it has not:
+ALTER TEXT SEARCH DICTIONARY intdict (MAXLEN = 6.7);
+select ts_lexize('intdict', '-40865854');
+ ts_lexize
+-----------
+ {-40865}
+(1 row)
+
+select ts_lexize('intdict', '+40865854');
+ ts_lexize
+-----------
+ {+40865}
+(1 row)
+
+ALTER TEXT SEARCH DICTIONARY intdict (ABSVAL = true);
+select ts_lexize('intdict', '-40865854');
+ ts_lexize
+-----------
+ {408658}
+(1 row)
+
+select ts_lexize('intdict', '+40865854');
+ ts_lexize
+-----------
+ {408658}
+(1 row)
+
+ALTER TEXT SEARCH DICTIONARY intdict (REJECTLONG = 1);
+select ts_lexize('intdict', '-40865854');
+ ts_lexize
+-----------
+ {}
+(1 row)
+
+select ts_lexize('intdict', '-4086585');
+ ts_lexize
+-----------
+ {}
+(1 row)
+
+select ts_lexize('intdict', '-408658');
+ ts_lexize
+-----------
+ {408658}
+(1 row)
+
+SELECT dictinitoption FROM pg_ts_dict WHERE dictname = 'intdict';
+ dictinitoption
+-----------------------------------------------
+ maxlen = 6.7, absval = 'true', rejectlong = 1
+(1 row)
+
diff --git a/contrib/dict_int/sql/dict_int.sql b/contrib/dict_int/sql/dict_int.sql
new file mode 100644
index 0000000..4e2543d
--- /dev/null
+++ b/contrib/dict_int/sql/dict_int.sql
@@ -0,0 +1,69 @@
+CREATE EXTENSION dict_int;
+
+--lexize
+select ts_lexize('intdict', '511673');
+select ts_lexize('intdict', '129');
+select ts_lexize('intdict', '40865854');
+select ts_lexize('intdict', '952');
+select ts_lexize('intdict', '654980341');
+select ts_lexize('intdict', '09810106');
+select ts_lexize('intdict', '14262713');
+select ts_lexize('intdict', '6532082986');
+select ts_lexize('intdict', '0150061');
+select ts_lexize('intdict', '7778');
+select ts_lexize('intdict', '9547');
+select ts_lexize('intdict', '753395478');
+select ts_lexize('intdict', '647652');
+select ts_lexize('intdict', '6988655574');
+select ts_lexize('intdict', '1279');
+select ts_lexize('intdict', '1266645909');
+select ts_lexize('intdict', '7594193969');
+select ts_lexize('intdict', '16928207');
+select ts_lexize('intdict', '196850350328');
+select ts_lexize('intdict', '22026985592');
+select ts_lexize('intdict', '2063765');
+select ts_lexize('intdict', '242387310');
+select ts_lexize('intdict', '93595');
+select ts_lexize('intdict', '9374');
+select ts_lexize('intdict', '996969');
+select ts_lexize('intdict', '353595982');
+select ts_lexize('intdict', '925860');
+select ts_lexize('intdict', '11848378337');
+select ts_lexize('intdict', '333');
+select ts_lexize('intdict', '799287416765');
+select ts_lexize('intdict', '745939');
+select ts_lexize('intdict', '67601305734');
+select ts_lexize('intdict', '3361113');
+select ts_lexize('intdict', '9033778607');
+select ts_lexize('intdict', '7507648');
+select ts_lexize('intdict', '1166');
+select ts_lexize('intdict', '9360498');
+select ts_lexize('intdict', '917795');
+select ts_lexize('intdict', '9387894');
+select ts_lexize('intdict', '42764329');
+select ts_lexize('intdict', '564062');
+select ts_lexize('intdict', '5413377');
+select ts_lexize('intdict', '060965');
+select ts_lexize('intdict', '08273593');
+select ts_lexize('intdict', '593556010144');
+select ts_lexize('intdict', '17988843352');
+select ts_lexize('intdict', '252281774');
+select ts_lexize('intdict', '313425');
+select ts_lexize('intdict', '641439323669');
+select ts_lexize('intdict', '314532610153');
+
+ALTER TEXT SEARCH DICTIONARY intdict (MAXLEN = -214783648); -- fail
+-- This ought to fail, perhaps, but historically it has not:
+ALTER TEXT SEARCH DICTIONARY intdict (MAXLEN = 6.7);
+
+select ts_lexize('intdict', '-40865854');
+select ts_lexize('intdict', '+40865854');
+ALTER TEXT SEARCH DICTIONARY intdict (ABSVAL = true);
+select ts_lexize('intdict', '-40865854');
+select ts_lexize('intdict', '+40865854');
+ALTER TEXT SEARCH DICTIONARY intdict (REJECTLONG = 1);
+select ts_lexize('intdict', '-40865854');
+select ts_lexize('intdict', '-4086585');
+select ts_lexize('intdict', '-408658');
+
+SELECT dictinitoption FROM pg_ts_dict WHERE dictname = 'intdict';
generated by cgit v1.2.3 (git 2.39.1) at 2024-11-30 21:08:49 +0000