Small fixes

Author: marcoslot

expected/time_interval.out

@@ -1,6 +1,7 @@
 create extension pg_incremental cascade;
 create schema time_range;
 set search_path to time_range;
+set client_min_messages to warning;
 -- create a source table
 create table events (
   event_id bigint generated always as identity,
@@ -29,7 +30,6 @@ select incremental.create_time_interval_pipeline('event-aggregation', '1 day',
   where event_time >= $1 and event_time < $2
   group by 1
   $$);
-NOTICE:  pipeline event-aggregation: processing time range from Fri Dec 31 16:00:00 1999 PST to Fri Dec 06 00:00:00 2024 PST
  create_time_interval_pipeline 
 -------------------------------
 
@@ -50,7 +50,6 @@ select sum(event_count) from events_agg;
 insert into events (client_id, path, response_time)
 select s, '/page-' || (s % 3), random() from generate_series(1,100) s;
 call incremental.execute_pipeline('event-aggregation');
-NOTICE:  pipeline event-aggregation: no rows to process
 select count(*) from events;
  count 
 -------
@@ -65,10 +64,4 @@ select sum(event_count) from events_agg;
 (1 row)
 
 drop schema time_range cascade;
-NOTICE:  drop cascades to 2 other objects
-DETAIL:  drop cascades to table events
-drop cascades to table events_agg
 drop extension pg_incremental;
-NOTICE:  drop cascades to 2 other objects
-DETAIL:  drop cascades to function incremental._drop_extension_trigger()
-drop cascades to event trigger incremental_drop_extension_trigger

pg_incremental--1.0.sql

@@ -52,7 +52,7 @@ GRANT SELECT ON incremental.processed_files TO public;
 
 CREATE FUNCTION incremental.create_sequence_pipeline(
     pipeline_name text,
-    sequence_name regclass,
+    source_table_name regclass,
     command text,
     schedule text default '* * * * *',
     execute_immediately bool default true)

sql/time_interval.sql

@@ -1,6 +1,7 @@
 create extension pg_incremental cascade;
 create schema time_range;
 set search_path to time_range;
+set client_min_messages to warning;
 
 -- create a source table
 create table events (

src/pipeline.c

@@ -50,7 +50,7 @@ incremental_create_sequence_pipeline(PG_FUNCTION_ARGS)
 	if (PG_ARGISNULL(0))
 		ereport(ERROR, (errmsg("pipeline_name cannot be NULL")));
 	if (PG_ARGISNULL(1))
-		ereport(ERROR, (errmsg("sequence_name cannot be NULL")));
+		ereport(ERROR, (errmsg("source_table_name cannot be NULL")));
 	if (PG_ARGISNULL(2))
 		ereport(ERROR, (errmsg("command cannot be NULL")));
 
@@ -64,14 +64,18 @@ incremental_create_sequence_pipeline(PG_FUNCTION_ARGS)
 
 	switch (get_rel_relkind(sequenceId))
 	{
+			/*
+			 * We allow source_table_name to be a sequence, and this is
+			 * necessary if there are multiple sequences.
+			 */
 		case RELKIND_SEQUENCE:
 			{
 				int32		columnNumber = 0;
 
 				if (!sequenceIsOwned(sequenceId, DEPENDENCY_AUTO, &sourceRelationId, &columnNumber))
 				{
 					ereport(ERROR, (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
-									errmsg("enly sequences that are owned by a table are supported")));
+									errmsg("only sequences that are owned by a table are supported")));
 				}
 
 				break;

src/sequence.c

@@ -319,17 +319,15 @@ FindSequenceForRelation(Oid relationId)
 		ereport(ERROR, (errcode(ERRCODE_UNDEFINED_OBJECT),
 						errmsg("relation \"%s\" does not have any sequences associated "
 							   "with it",
-							   get_rel_name(relationId)),
-						errhint("Specify the name of the sequence to use for the "
-								"pipeline as the argument")));
+							   get_rel_name(relationId))));
 
 	if (list_length(sequences) > 1)
 		ereport(ERROR, (errcode(ERRCODE_UNDEFINED_OBJECT),
 						errmsg("relation \"%s\" has multiple sequences associated "
 							   "with it",
 							   get_rel_name(relationId)),
-						errhint("Specify the name of the sequence to use for the "
-								"pipeline as the argument")));
+						errhint("Specify the name of the sequence to use instead of "
+								"the table name")));
 
 	return linitial_oid(sequences);
 }