>
> Are you using INSERT or COPY to insert your data? COPY
> tends to be a lot faster than separate INSERTs, especially
> if you don't wrap the INSERTs in a transaction block and
> COMMIT them in batches.
>
But I do not use File or Stdio. The data is coming in row by row real-time. I need to insert the data programmatic real-time into the database. That's why I use INSERT. But maybe I miss out something on the usage of COPY, please advice :)
> As an aside, I hope you do realise that your primary key
> isn't guaranteed to be unique across your child tables? The
> reason is the same one that you already quoted for indexes
> spanning multiple tables - a primary key is implemented
> using a unique index after all.
>
> If that matters, what you can do is make your primary key a
> foreign key to a new table with just the primary key column
> in it. Make sure you always insert a record in the primary
> key table along with the one referencing it, so that you
> will get a unique violation when you try to insert a record
> for which the primary key already exists. This will of
> course slow things down some, but if it's necessary that's
> the price to pay.
>
Oh. I didn't notice that. Thanks for pointing out. Luckily, thanks God. It doesn't matter much at this moment :)
--
Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
