Allow to use experimental functions in caggs

Current implementation of caggs can't find a bucketing function
if it's declared in the experimental schema. This patch fixes it.

Also the patch adds `debug_notice` test to IGNORE list on AppVeyor.
The corresponding test generates an extra "DEBUG: rehashing catalog
cache" message which is not critical. It seems to be stable on Linux.

appveyor.yml

@@ -203,7 +203,7 @@ test_script:
     # killer. Therefore, we need to ignore the results of the
     # remote_connection and remote_txn tests.
 
-    docker exec -e IGNORES="bgw_db_scheduler compression_algos continuous_aggs_bgw ordered_append_join-12 remote_connection remote_txn " -e SKIPS="bgw_db_scheduler" -e TEST_TABLESPACE1_PREFIX="C:\Users\$env:UserName\Documents\tablespace1\" -e TEST_TABLESPACE2_PREFIX="C:\Users\$env:UserName\Documents\tablespace2\" -e TEST_SPINWAIT_ITERS=10000 -e USER=postgres -it pgregress /bin/bash -c "cd /timescaledb/build && make -k regresschecklocal-t regresschecklocal-shared"
+    docker exec -e IGNORES="bgw_db_scheduler compression_algos continuous_aggs_bgw debug_notice ordered_append_join-12 remote_connection remote_txn " -e SKIPS="bgw_db_scheduler" -e TEST_TABLESPACE1_PREFIX="C:\Users\$env:UserName\Documents\tablespace1\" -e TEST_TABLESPACE2_PREFIX="C:\Users\$env:UserName\Documents\tablespace2\" -e TEST_SPINWAIT_ITERS=10000 -e USER=postgres -it pgregress /bin/bash -c "cd /timescaledb/build && make -k regresschecklocal-t regresschecklocal-shared"
 
     if( -not $? -or -not $TESTS1 ) { exit 1 }
 

src/extension.c

@@ -31,6 +31,13 @@
 
 #define TS_UPDATE_SCRIPT_CONFIG_VAR "timescaledb.update_script_stage"
 #define POST_UPDATE "post"
+/*
+ * The name of the experimental schema.
+ *
+ * Call ts_extension_schema_name() or ts_experimental_schema_name() for
+ * consistency. Don't use this macro directly.
+ */
+#define TS_EXPERIMENTAL_SCHEMA_NAME "timescaledb_experimental"
 static Oid extension_proxy_oid = InvalidOid;
 
 /*
@@ -202,6 +209,12 @@ ts_extension_schema_name(void)
 	return get_namespace_name(ts_extension_schema_oid());
 }
 
+const char *
+ts_experimental_schema_name(void)
+{
+	return TS_EXPERIMENTAL_SCHEMA_NAME;
+}
+
 /*
  *	Called upon all Relcache invalidate events.
  *	Returns whether or not to invalidate the entire extension.

src/extension.h

@@ -15,6 +15,7 @@ extern void ts_extension_check_version(const char *so_version);
 extern void ts_extension_check_server_version(void);
 extern Oid ts_extension_schema_oid(void);
 extern TSDLLEXPORT char *ts_extension_schema_name(void);
+extern const char *ts_experimental_schema_name(void);
 extern const char *ts_extension_get_so_name(void);
 extern TSDLLEXPORT const char *ts_extension_get_version(void);
 

src/func_cache.c

@@ -176,7 +176,7 @@ typedef struct FuncEntry
 /* Information about functions that we put in the cache */
 static FuncInfo funcinfo[] = {
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket",
 		.nargs = 2,
@@ -185,7 +185,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket",
 		.nargs = 3,
@@ -194,7 +194,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket",
 		.nargs = 2,
@@ -203,7 +203,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket",
 		.nargs = 3,
@@ -212,7 +212,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket",
 		.nargs = 2,
@@ -221,7 +221,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket",
 		.nargs = 3,
@@ -230,7 +230,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket",
 		.nargs = 2,
@@ -239,7 +239,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket",
 		.nargs = 3,
@@ -248,7 +248,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket",
 		.nargs = 2,
@@ -257,7 +257,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket",
 		.nargs = 3,
@@ -266,7 +266,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket",
 		.nargs = 2,
@@ -275,7 +275,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket",
 		.nargs = 3,
@@ -283,9 +283,44 @@ static FuncInfo funcinfo[] = {
 		.group_estimate = time_bucket_group_estimate,
 		.sort_transform = time_bucket_sort_transform,
 	},
-
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE_EXPERIMENTAL,
+		.is_bucketing_func = true,
+		.funcname = "time_bucket_ng",
+		.nargs = 2,
+		.arg_types = { INTERVALOID, DATEOID },
+		.group_estimate = time_bucket_group_estimate,
+		.sort_transform = time_bucket_sort_transform,
+	},
+	{
+		.origin = ORIGIN_TIMESCALE_EXPERIMENTAL,
+		.is_bucketing_func = true,
+		.funcname = "time_bucket_ng",
+		.nargs = 3,
+		.arg_types = { INTERVALOID, DATEOID, DATEOID },
+		.group_estimate = time_bucket_group_estimate,
+		.sort_transform = time_bucket_sort_transform,
+	},
+	{
+		.origin = ORIGIN_TIMESCALE_EXPERIMENTAL,
+		.is_bucketing_func = true,
+		.funcname = "time_bucket_ng",
+		.nargs = 2,
+		.arg_types = { INTERVALOID, TIMESTAMPOID },
+		.group_estimate = time_bucket_group_estimate,
+		.sort_transform = time_bucket_sort_transform,
+	},
+	{
+		.origin = ORIGIN_TIMESCALE_EXPERIMENTAL,
+		.is_bucketing_func = true,
+		.funcname = "time_bucket_ng",
+		.nargs = 3,
+		.arg_types = { INTERVALOID, TIMESTAMPOID, TIMESTAMPOID },
+		.group_estimate = time_bucket_group_estimate,
+		.sort_transform = time_bucket_sort_transform,
+	},
+	{
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket_gapfill",
 		.nargs = 4,
@@ -294,7 +329,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_gapfill_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket_gapfill",
 		.nargs = 4,
@@ -303,7 +338,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_gapfill_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket_gapfill",
 		.nargs = 4,
@@ -312,7 +347,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_gapfill_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket_gapfill",
 		.nargs = 4,
@@ -321,7 +356,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_gapfill_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket_gapfill",
 		.nargs = 4,
@@ -330,7 +365,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = time_bucket_gapfill_sort_transform,
 	},
 	{
-		.is_timescaledb_func = true,
+		.origin = ORIGIN_TIMESCALE,
 		.is_bucketing_func = true,
 		.funcname = "time_bucket_gapfill",
 		.nargs = 4,
@@ -340,7 +375,7 @@ static FuncInfo funcinfo[] = {
 	},
 
 	{
-		.is_timescaledb_func = false,
+		.origin = ORIGIN_POSTGRES,
 		.is_bucketing_func = true,
 		.funcname = "date_trunc",
 		.nargs = 2,
@@ -349,7 +384,7 @@ static FuncInfo funcinfo[] = {
 		.sort_transform = date_trunc_sort_transform,
 	},
 	{
-		.is_timescaledb_func = false,
+		.origin = ORIGIN_POSTGRES,
 		.is_bucketing_func = true,
 		.funcname = "date_trunc",
 		.nargs = 2,
@@ -379,6 +414,7 @@ initialize_func_info()
 		.hcxt = CacheMemoryContext,
 	};
 	Oid extension_nsp = ts_extension_schema_oid();
+	Oid experimental_nsp = get_namespace_oid(ts_experimental_schema_name(), false);
 	Oid pg_nsp = get_namespace_oid("pg_catalog", false);
 	HeapTuple tuple;
 	Relation rel;
@@ -391,12 +427,21 @@ initialize_func_info()
 	for (i = 0; i < _MAX_CACHE_FUNCTIONS; i++)
 	{
 		FuncInfo *finfo = &funcinfo[i];
-		Oid namespaceoid = finfo->is_timescaledb_func ? extension_nsp : pg_nsp;
+		Oid namespaceoid = pg_nsp;
 		oidvector *paramtypes = buildoidvector(finfo->arg_types, finfo->nargs);
 		FuncEntry *fentry;
 		bool hash_found;
 		Oid funcid;
 
+		if (finfo->origin == ORIGIN_TIMESCALE)
+		{
+			namespaceoid = extension_nsp;
+		}
+		else if (finfo->origin == ORIGIN_TIMESCALE_EXPERIMENTAL)
+		{
+			namespaceoid = experimental_nsp;
+		}
+
 		tuple = SearchSysCache3(PROCNAMEARGSNSP,
 								PointerGetDatum(finfo->funcname),
 								PointerGetDatum(paramtypes),

src/func_cache.h

@@ -16,10 +16,28 @@
 typedef Expr *(*sort_transform_func)(FuncExpr *func);
 typedef double (*group_estimate_func)(PlannerInfo *root, FuncExpr *expr, double path_rows);
 
+/* Describes the function origin */
+typedef enum
+{
+	/*
+	 * Function is provided by PostgreSQL.
+	 */
+	ORIGIN_POSTGRES = 0,
+	/*
+	 * Function is provided by TimescaleDB.
+	 */
+	ORIGIN_TIMESCALE = 1,
+	/*
+	 * Fuction is provided by TimescaleDB and is experimental.
+	 * It should be looked for in the experimental schema.
+	 */
+	ORIGIN_TIMESCALE_EXPERIMENTAL = 2,
+} FuncOrigin;
+
 typedef struct FuncInfo
 {
 	const char *funcname;
-	bool is_timescaledb_func;
+	FuncOrigin origin;
 	bool is_bucketing_func;
 	int nargs;
 	Oid arg_types[FUNC_CACHE_MAX_FUNC_ARGS];

tsl/src/continuous_aggs/create.c

@@ -1069,9 +1069,18 @@ get_partialize_funcexpr(Aggref *agg)
 static bool
 is_valid_bucketing_function(Oid funcid)
 {
+	bool is_timescale;
 	FuncInfo *finfo = ts_func_cache_get_bucketing_func(funcid);
 
-	return finfo != NULL && finfo->is_timescaledb_func && finfo->nargs == 2;
+	if (finfo == NULL)
+	{
+		return false;
+	}
+
+	is_timescale =
+		(finfo->origin == ORIGIN_TIMESCALE) || (finfo->origin == ORIGIN_TIMESCALE_EXPERIMENTAL);
+
+	return is_timescale && (finfo->nargs == 2);
 }
 
 /*initialize MatTableColumnInfo */

tsl/test/expected/continuous_aggs_experimental.out

@@ -0,0 +1,55 @@
+-- This file and its contents are licensed under the Timescale License.
+-- Please see the included NOTICE for copyright information and
+-- LICENSE-TIMESCALE for a copy of the license.
+-- Make sure experimental immutable function with 2 arguments can be used in caggs.
+-- Functions with 3 arguments and/or stable functions are currently not supported in caggs.
+CREATE TABLE conditions(
+  day DATE NOT NULL,
+  city text NOT NULL,
+  temperature INT NOT NULL);
+SELECT create_hypertable(
+  'conditions', 'day',
+  chunk_time_interval => INTERVAL '1 day'
+);
+    create_hypertable    
+-------------------------
+ (1,public,conditions,t)
+(1 row)
+
+INSERT INTO conditions (day, city, temperature) VALUES
+  ('2021-06-14', 'Moscow', 26),
+  ('2021-06-15', 'Moscow', 22),
+  ('2021-06-16', 'Moscow', 24),
+  ('2021-06-17', 'Moscow', 24),
+  ('2021-06-18', 'Moscow', 27),
+  ('2021-06-19', 'Moscow', 28),
+  ('2021-06-20', 'Moscow', 30),
+  ('2021-06-21', 'Moscow', 31),
+  ('2021-06-22', 'Moscow', 34),
+  ('2021-06-23', 'Moscow', 34),
+  ('2021-06-24', 'Moscow', 34),
+  ('2021-06-25', 'Moscow', 32),
+  ('2021-06-26', 'Moscow', 32),
+  ('2021-06-27', 'Moscow', 31);
+CREATE MATERIALIZED VIEW conditions_summary_weekly
+WITH (timescaledb.continuous) AS
+SELECT city,
+       timescaledb_experimental.time_bucket_ng('7 days', day) AS bucket,
+       MIN(temperature),
+       MAX(temperature)
+FROM conditions
+GROUP BY city, bucket;
+NOTICE:  refreshing continuous aggregate "conditions_summary_weekly"
+SELECT to_char(bucket, 'YYYY-MM-DD'), city, min, max
+FROM conditions_summary_weekly
+ORDER BY bucket;
+  to_char   |  city  | min | max 
+------------+--------+-----+-----
+ 2021-06-12 | Moscow |  22 |  27
+ 2021-06-19 | Moscow |  28 |  34
+ 2021-06-26 | Moscow |  31 |  32
+(3 rows)
+
+DROP TABLE conditions CASCADE;
+NOTICE:  drop cascades to 3 other objects
+NOTICE:  drop cascades to 2 other objects

tsl/test/sql/CMakeLists.txt

@@ -10,6 +10,7 @@ set(TEST_FILES
     continuous_aggs_policy.sql
     continuous_aggs_refresh.sql
     continuous_aggs_watermark.sql
+    continuous_aggs_experimental.sql
     dist_views.sql
     move.sql
     partialize_finalize.sql

tsl/test/sql/continuous_aggs_experimental.sql

@@ -0,0 +1,47 @@
+-- This file and its contents are licensed under the Timescale License.
+-- Please see the included NOTICE for copyright information and
+-- LICENSE-TIMESCALE for a copy of the license.
+
+-- Make sure experimental immutable function with 2 arguments can be used in caggs.
+-- Functions with 3 arguments and/or stable functions are currently not supported in caggs.
+
+CREATE TABLE conditions(
+  day DATE NOT NULL,
+  city text NOT NULL,
+  temperature INT NOT NULL);
+
+SELECT create_hypertable(
+  'conditions', 'day',
+  chunk_time_interval => INTERVAL '1 day'
+);
+
+INSERT INTO conditions (day, city, temperature) VALUES
+  ('2021-06-14', 'Moscow', 26),
+  ('2021-06-15', 'Moscow', 22),
+  ('2021-06-16', 'Moscow', 24),
+  ('2021-06-17', 'Moscow', 24),
+  ('2021-06-18', 'Moscow', 27),
+  ('2021-06-19', 'Moscow', 28),
+  ('2021-06-20', 'Moscow', 30),
+  ('2021-06-21', 'Moscow', 31),
+  ('2021-06-22', 'Moscow', 34),
+  ('2021-06-23', 'Moscow', 34),
+  ('2021-06-24', 'Moscow', 34),
+  ('2021-06-25', 'Moscow', 32),
+  ('2021-06-26', 'Moscow', 32),
+  ('2021-06-27', 'Moscow', 31);
+
+CREATE MATERIALIZED VIEW conditions_summary_weekly
+WITH (timescaledb.continuous) AS
+SELECT city,
+       timescaledb_experimental.time_bucket_ng('7 days', day) AS bucket,
+       MIN(temperature),
+       MAX(temperature)
+FROM conditions
+GROUP BY city, bucket;
+
+SELECT to_char(bucket, 'YYYY-MM-DD'), city, min, max
+FROM conditions_summary_weekly
+ORDER BY bucket;
+
+DROP TABLE conditions CASCADE;