Re: Updates on updates

First thanks for your message, highly interesting (despite its title ;-)

Shachar Shemesh <psql(at)shemesh(dot)biz> writes:

> Things are slightly worse, in fact. Because the ordering of the records
> in a dynaset are guarenteed to remain the same as when they were first
> fetched,

PostgreSQL transaction isolation implementation is based on snapshots,
so this looks sensible. Except I don't get the meaning of: "the
ordering" (so i just suppressed it from the sentence in order to
understand it :-)

But what happens if you lower the transaction isolation level to "read
committed"? do the assertions above still hold? Do later fetches
still see old data? If yes, is it because of the way the
driver/protocol is implemented or because of the engine itself? (the
latter would be surprising).