Functions and Commits

[pg1234]

I have been reading through all the threads and trying a few different things. Essentially, I have a function that loops through a table and then performs some updates/inserts against other tables based on the row from the FOR LOOP.

Currently, none of these inserts/updates are visible to other users until the LOOP and function are complete (i.e. the end of the function is the commit).
However, I would like to make these insert/updates visible as soon as a row in the for loop is complete. Is this possible? or is this the problem of a nested transaction?

I am about to try dblink. Is there anyway to accomplish this?

[michaelb]

Nested transactions are (kind of) supported in the latest versions of Postgres.
This involves usage of savepoints and I think it may have a negative impact on performance when you processing a big volume of data.
If you want to try it check out read this

It is quite common to make updates available when entire transaction is complete.
I wonder why are you trying to do this, does this update takes a very long time, or is there some other reason?
Perhaps it would help if you describe the problem before committing to a specific solution, like dblink, or nested transactions.