Re: CREATE TABLE LIKE INCLUDING TRIGGERS

On Mon, Sep 29, 2025 at 5:35 PM jian he <jian.universality@gmail.com> wrote:
>
> hi.
>
> poc demo:
> CREATE TABLE main_table (a int, b int);
> CREATE FUNCTION trigger_func() RETURNS trigger LANGUAGE plpgsql AS '
> BEGIN
>     RAISE NOTICE ''trigger_func(%) called: action = %, when = %, level = %'',
>                   TG_ARGV[0], TG_OP, TG_WHEN, TG_LEVEL;
>     RETURN NULL;
> END;';
> CREATE TRIGGER modified_a BEFORE UPDATE OF a ON main_table
> FOR EACH ROW WHEN (OLD.a <> NEW.a) EXECUTE PROCEDURE trigger_func('modified_a');
>
> CREATE TABLE main_table1(LIKE main_table INCLUDING TRIGGERS INCLUDING COMMENTS);
> \d main_table1
>             Table "public.main_table1"
>  Column |  Type   | Collation | Nullable | Default
> --------+---------+-----------+----------+---------
>  a      | integer |           |          |
>  b      | integer |           |          |
> Triggers:
>     modified_a BEFORE UPDATE OF a ON main_table1 FOR EACH ROW WHEN
> (old.a <> new.a) EXECUTE FUNCTION trigger_func('modified_a')
>
> foreign key associated internal triggers won't be copied to the new table.
> source table trigger associated comment will be copied to the new table,
> if INCLUDING COMMENTS is specified.
>

per
https://api.cirrus-ci.com/v1/artifact/task/5194724417470464/testrun/build/testrun/regress/regress/regression.diffs

there are many table name as t1 in regress test, add tests like
""CREATE TABLE t1 (a int, b text, c int);"
may result in error
+ERROR:  relation "t1" already exists
in some OS.
So I changed the table name to avoid parallel regess test failure.