On Tuesday, January 11, 2022, <
benj.dev@laposte.net> wrote:
SHOW search_path -- => "$user", public;
DROP TABLE IF EXISTS my_table;
-- Executing each row on differents transactions but in the same session
/*Session A - Transaction 1*/ SELECT * FROM test_search_path_v1(true); -- => OK takes table from pg_temp (no existing table in public)
/*Session A - Transaction 2*/ SELECT * FROM test_search_path_v1(false); -- => OK takes table from public
/*Session A - Transaction 3*/ SELECT * FROM test_search_path_v1(true); -- => OK takes table from pg_temp (and the existing from public)
/*Session A - Transaction 4*/ SELECT * FROM test_search_path_v1(false); -- => OK takes table from public
/*Session A - Transaction 5*/ SELECT * FROM test_search_path_v1(true); -- => NOK => it takes public and not pg_temp
Per [1], you are violating:
“Because PL/pgSQL saves prepared statements and sometimes execution plans in this way, SQL commands that appear directly in a PL/pgSQL function must refer to the same tables and columns on every execution; that is, you cannot use a parameter as the name of a table or column in an SQL command.”
While that wording is odd, it is not even possible to use variables in place of table and column names, what you are doing here is effectively the same. I cannot explain the precise observed behavior, and maybe there is a bug here, but on its face what you are doing is simply invalid in face of how the session relation cache works.
David J.
