What is the benefit of embedding pg_advisory_lock inside SELECT vs explicitly using another statement?

Question

I cannot really understand the example from https://www.postgresql.org/docs/9.1/static/explicit-locking.html

where they embed the lock along with the SELECT clause SELECT pg_advisory_lock(id) FROM foo WHERE id = 12345; -- ok SELECT pg_advisory_lock(id) FROM foo WHERE id = 12345; -- ok

What is it selecting from FOO ? I would understand better if it is like

SELECT pg_advisory_lock(123); //lock
SELECT * FROM foo WHERE id = 12345;

where it is explicitly locking the block. I can't seem to find explanation on how to really use advisory lock anywhere that explains the differnece between both Embedding and Explicitly on its own statement.

Answer 1

Doing SELECT pg_advisory_lock(123); will create a lock on 123 , whether it is a valid value or not. Doing SELECT pg_advisory_lock(id) FROM foo WHERE id = 123; will create the lock only if there is an entry for ID 123 in table foo.

Let's note the line found in the pg_locks doc :

The actual meaning of the keys is up to the user

which tends to imply that the select/from/where syntax is for associating the lock to an existing row, while the lone select syntax is for a broader meaning, like an application-wide lock.