Re: How to speed up min/max(id) in 50M rows table?

> It's possible to combine independent indexes for resolving AND-type
> queries, but the combination process does not preserve ordering, so
> it's useless for this type of situation.

Ok, I'm going to try the double column index. Your suggestion about the index with nulls left out worked great btw. Min/Max is instantly now.

I figure though that the double column index would not work for queries like:

select min(time) from transactions where payment_id is null

So for that situation I tried whether a specific index helped, i.e. :

create index transactions__time_payment_id__null__idx on transactions(time) where payment_id is null;

But this does not really seem to help. It might be better to see if I can refactor the DB design though to not use nulls.

Regards

_________________________________________________________________
Express yourself instantly with MSN Messenger! Download today it's FREE!
http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/