Merge pull request #86 from cybertec-postgresql/smoc_gin_options

Configurable smoc_gin_ops index resolution

Author: vitcpp

Makefile

@@ -99,6 +99,7 @@ healpix_bare/healpix_bare.o : healpix_bare/healpix_bare.c
 
 pg_version := $(word 2,$(shell $(PG_CONFIG) --version))
 has_support_functions = $(if $(filter-out 9.% 10.% 11.%,$(pg_version)),y,n)
+has_index_options = $(if $(filter-out 9.% 10.% 11.% 12.%,$(pg_version)),y,n)
 
 crushtest: TESTS += $(CRUSH_TESTS)
 crushtest: installcheck
@@ -108,6 +109,13 @@ PGS_SQL    += pgs_gist_support.sql
 TESTS      += gist_support
 endif
 
+ifneq ($(USE_HEALPIX),0)
+ifeq ($(has_index_options),y)
+PGS_SQL    += pgs_moc_options.sql
+TESTS      += moc_options
+endif
+endif
+
 # "make test" uses a special initialization file that doesn't rely on "create extension"
 test: pg_sphere.test.sql
 	$(pg_regress_installcheck) --temp-instance=tmp_check $(REGRESS_OPTS) init_test $(TESTS)
@@ -186,6 +194,11 @@ pg_sphere--1.3.0--1.3.1.sql:
 ifeq ($(has_support_functions),y)
 pg_sphere--1.3.1--1.3.2.sql: pgs_gist_support.sql.in
 endif
+ifneq ($(USE_HEALPIX),0)
+ifeq ($(has_index_options),y)
+pg_sphere--1.3.1--1.3.2.sql: pgs_moc_options.sql.in
+endif
+endif
 pg_sphere--1.3.1--1.3.2.sql: pgs_circle_sel.sql.in
 	cat upgrade_scripts/$@.in $^ > $@
 

doc/indices.sgm

@@ -121,23 +121,32 @@
           The index works by casting all contained smocs to a fixed level, and
           for each pixel at that level, storing which smocs overlap with that
           pixel. This is especially beneficial for "overlaps" queries using
-          the <literal>&amp;&amp;</literal> operator. Two levels of granularity
-          are provided: the default opclass <literal>smoc_gin_ops</literal>
-          works on level 5 with a resolution of 12288 pixels, while the
-          opclass <literal>smoc_gin_ops_fine</literal> works on level 8 with
-          786432 pixels. The downside of that approach is that storing large
+          the <literal>&amp;&amp;</literal> operator.
+          The downside of that approach is that storing large
           smocs like "all sky" (<literal>0/0-11</literal>) produces a large
           number of index entries.
         </para>
+        <para>
+          The default opclass <literal>smoc_gin_ops</literal> defaults to
+          working on level 5 with a resolution of 12288 pixels (12 * 4^5).
+          An alternative granularity can be selected by setting the
+          <literal>order</literal> parameter on the opclass (integer value
+          between 0 and 12; option only available on PG 13 and later).
+          The alternative <literal>smoc_gin_ops_fine</literal> opclass works
+          on level 8 with 786432 pixels.
+        </para>
         <example>
           <title>Index of smoc coverage objects</title>
           <programlisting>
 <![CDATA[CREATE TABLE ivoa (]]>
 <![CDATA[  coverage smoc NOT NULL]]>
 <![CDATA[);]]>
 <![CDATA[-- Put in data now]]>
+<![CDATA[-- Create index with the defaut smoc_gin_ops opclass (order 5)]]>
 <![CDATA[CREATE INDEX ON ivoa USING GIN (coverage);]]>
-<![CDATA[-- Alternative index with more detail]]>
+<![CDATA[-- Alternative index with more detail on order 7]]>
+<![CDATA[CREATE INDEX ivoa_order_7_idx ON ivoa USING GIN (coverage smoc_gin_ops (order = 7));]]>
+<![CDATA[-- Alternative operator class with fixed order 8]]>
 <![CDATA[CREATE INDEX ivoa_fine_idx ON ivoa USING GIN (coverage smoc_gin_ops_fine);]]>
           </programlisting>
         </example>

expected/moc_options.out

@@ -0,0 +1,40 @@
+create table moc_opt (m smoc);
+insert into moc_opt select format('9/%s', i)::smoc from generate_series(1, 1000) g(i);
+analyze moc_opt;
+create index moc_opt5 on moc_opt using gin (m);
+explain (analyze, costs off, timing off, summary off) select * from moc_opt where m && '9/1';
+                          QUERY PLAN                           
+---------------------------------------------------------------
+ Bitmap Heap Scan on moc_opt (actual rows=1 loops=1)
+   Recheck Cond: (m && '9/1'::smoc)
+   Rows Removed by Index Recheck: 254
+   Heap Blocks: exact=4
+   ->  Bitmap Index Scan on moc_opt5 (actual rows=255 loops=1)
+         Index Cond: (m && '9/1'::smoc)
+(6 rows)
+
+drop index moc_opt5;
+create index moc_opt8 on moc_opt using gin (m smoc_gin_ops_fine);
+explain (analyze, costs off, timing off, summary off) select * from moc_opt where m && '9/1';
+                         QUERY PLAN                          
+-------------------------------------------------------------
+ Bitmap Heap Scan on moc_opt (actual rows=1 loops=1)
+   Recheck Cond: (m && '9/1'::smoc)
+   Rows Removed by Index Recheck: 2
+   Heap Blocks: exact=1
+   ->  Bitmap Index Scan on moc_opt8 (actual rows=3 loops=1)
+         Index Cond: (m && '9/1'::smoc)
+(6 rows)
+
+drop index moc_opt8;
+create index moc_opt9 on moc_opt using gin (m smoc_gin_ops (order = 9));
+explain (analyze, costs off, timing off, summary off) select * from moc_opt where m && '9/1';
+                         QUERY PLAN                          
+-------------------------------------------------------------
+ Bitmap Heap Scan on moc_opt (actual rows=1 loops=1)
+   Recheck Cond: (m && '9/1'::smoc)
+   Heap Blocks: exact=1
+   ->  Bitmap Index Scan on moc_opt9 (actual rows=1 loops=1)
+         Index Cond: (m && '9/1'::smoc)
+(5 rows)
+

pgs_moc_ops.sql.in

@@ -354,6 +354,7 @@ CREATE OPERATOR CLASS smoc_gin_ops
 		FUNCTION 4 smoc_gin_consistent (internal, int2, smoc, int4, internal, internal, internal, internal),
 		--FUNCTION 5 smoc_gin_compare_partial (),
 		--FUNCTION 6 smoc_gin_tri_consistent (),
+		--FUNCTION 7 (smoc) smoc_gin_options (internal), -- needs PG13
 		STORAGE int4;
 
 CREATE OPERATOR CLASS smoc_gin_ops_fine

pgs_moc_options.sql.in

@@ -0,0 +1,12 @@
+-- GIN opclass options
+
+CREATE FUNCTION smoc_gin_options (internal)
+	RETURNS void
+	AS 'MODULE_PATHNAME'
+	LANGUAGE C
+	PARALLEL SAFE
+	IMMUTABLE
+	STRICT;
+
+ALTER OPERATOR FAMILY smoc_gin_ops USING gin
+	ADD FUNCTION 7 (smoc) smoc_gin_options (internal);

sql/moc_options.sql

@@ -0,0 +1,14 @@
+create table moc_opt (m smoc);
+insert into moc_opt select format('9/%s', i)::smoc from generate_series(1, 1000) g(i);
+analyze moc_opt;
+
+create index moc_opt5 on moc_opt using gin (m);
+explain (analyze, costs off, timing off, summary off) select * from moc_opt where m && '9/1';
+drop index moc_opt5;
+
+create index moc_opt8 on moc_opt using gin (m smoc_gin_ops_fine);
+explain (analyze, costs off, timing off, summary off) select * from moc_opt where m && '9/1';
+drop index moc_opt8;
+
+create index moc_opt9 on moc_opt using gin (m smoc_gin_ops (order = 9));
+explain (analyze, costs off, timing off, summary off) select * from moc_opt where m && '9/1';

src/moc.c

@@ -3,7 +3,8 @@
 
 #include <stddef.h>
 #include <string.h>
-#include <access/gin.h>
+#include "access/gin.h"
+#include "access/reloptions.h"
 
 #include "circle.h"
 #include "polygon.h"
@@ -45,6 +46,7 @@ PG_FUNCTION_INFO_V1(smoc_gin_extract_value_fine);
 PG_FUNCTION_INFO_V1(smoc_gin_extract_query);
 PG_FUNCTION_INFO_V1(smoc_gin_extract_query_fine);
 PG_FUNCTION_INFO_V1(smoc_gin_consistent);
+PG_FUNCTION_INFO_V1(smoc_gin_options);
 
 int32 smoc_output_type = 0;
 
@@ -1079,7 +1081,6 @@ smoc_gin_extract_internal(Smoc *moc_a, int32 *nkeys, int gin_order)
 			if (*nkeys >= nalloc)
 			{
 				nalloc *= 2;
-				Assert(nalloc < 2000000);
 				keys = repalloc(keys, nalloc * sizeof(Datum));
 			}
 			keys[(*nkeys)++] = Int32GetDatum(p);
@@ -1094,8 +1095,9 @@ smoc_gin_extract_value(PG_FUNCTION_ARGS)
 {
 	Smoc*	moc_a = (Smoc *) PG_DETOAST_DATUM(PG_GETARG_DATUM(0));
 	int32*	nkeys = (int32 *) PG_GETARG_POINTER(1);
+	int		order = SMOC_GIN_GET_ORDER();
 
-	PG_RETURN_DATUM(smoc_gin_extract_internal(moc_a, nkeys, MOC_GIN_ORDER));
+	PG_RETURN_DATUM(smoc_gin_extract_internal(moc_a, nkeys, order));
 }
 
 Datum
@@ -1114,13 +1116,14 @@ smoc_gin_extract_query(PG_FUNCTION_ARGS)
 	int32*	nkeys = (int32 *) PG_GETARG_POINTER(1);
 	StrategyNumber st = PG_GETARG_UINT16(2);
 	int32*	searchmode = (int32 *) PG_GETARG_POINTER(6);
+	int		order = SMOC_GIN_GET_ORDER();
 
 	if (st == MOC_GIN_STRATEGY_SUBSET || (st == MOC_GIN_STRATEGY_EQUAL && moc_a->area == 0))
 		*searchmode = GIN_SEARCH_MODE_INCLUDE_EMPTY;
 	else if (st == MOC_GIN_STRATEGY_UNEQUAL)
 		*searchmode = GIN_SEARCH_MODE_ALL;
 
-	PG_RETURN_DATUM(smoc_gin_extract_internal(moc_a, nkeys, MOC_GIN_ORDER));
+	PG_RETURN_DATUM(smoc_gin_extract_internal(moc_a, nkeys, order));
 }
 
 Datum
@@ -1202,3 +1205,21 @@ smoc_gin_consistent(PG_FUNCTION_ARGS)
 	/* not reached */
 	PG_RETURN_NULL();
 }
+
+#if PG_VERSION_NUM >= 130000
+Datum
+smoc_gin_options(PG_FUNCTION_ARGS)
+{
+	local_relopts *relopts = (local_relopts *) PG_GETARG_POINTER(0);
+
+	init_local_reloptions(relopts, sizeof(SMocGinOptions));
+	add_local_int_reloption(relopts, "order",
+							"smoc order to store in index",
+							MOC_GIN_ORDER_DEFAULT,
+							0,
+							12, /* maximum order fitting into 32bit */
+							offsetof(SMocGinOptions, order));
+
+	PG_RETURN_VOID();
+}
+#endif

src/pgs_moc.h

@@ -122,12 +122,29 @@ next_interval(int32 a)
 
 #define MOC_AREA_ALL_SKY 3458764513820540928
 
-#define MOC_GIN_ORDER 5 /* order 5 has 12 * 4^5 = 12288 pixels */
+#define MOC_GIN_ORDER_DEFAULT 5 /* order 5 has 12 * 4^5 = 12288 pixels */
 #define MOC_GIN_ORDER_FINE 8 /* order 8 has 12 * 4^8 = 786432 pixels */
 #define MOC_GIN_STRATEGY_INTERSECTS	1
 #define MOC_GIN_STRATEGY_SUBSET		2
 #define MOC_GIN_STRATEGY_SUPERSET	3
 #define MOC_GIN_STRATEGY_EQUAL		4
 #define MOC_GIN_STRATEGY_UNEQUAL	5
 
+/* smoc_gin_ops opclass options */
+#if PG_VERSION_NUM >= 130000
+Datum smoc_gin_options(PG_FUNCTION_ARGS);
+
+typedef struct
+{
+    int32       vl_len_;        /* varlena header (do not touch directly!) */
+    int         order;          /* smoc order to store in index (default 5) */
+} SMocGinOptions;
+
+#define SMOC_GIN_GET_ORDER()	(PG_HAS_OPCLASS_OPTIONS() ? \
+								 ((SMocGinOptions *) PG_GET_OPCLASS_OPTIONS())->order : \
+								 MOC_GIN_ORDER_DEFAULT)
+#else
+#define SMOC_GIN_GET_ORDER()	MOC_GIN_ORDER_DEFAULT
+#endif
+
 #endif