I'm importing several gigantic csv files into a table - so I'm wondering what would happen if I ran COPY FROM jobs concurrently?
2 Answers
If you're trying to speed things up, this is generally a good idea. I say "generally" because you can also speed things up by creating the table and loading data into in a single transaction, which reduces transaction logging. You'd have to test to see which is faster.
In PostgreSQL versions prior to 9.2 (currently in beta), the maximum number of concurrent beneficial COPYs tends to be 3. In 9.2 that goes up to at least 8, assuming you have the cores and I/O available.
-
9Not that I doubt your word, but can you provide a source on why "the maximum number of concurrent beneficial COPYs tends to be 3"? Commented Jul 12, 2012 at 14:19