What statistics do you have on the data? I suppose most messages are read
by low number of users, mostly 0 or one.
I can see two options to consider:
1) Use arrays to store information on which users have already read the
message. You may need GIN/GIST index to search fast.
2) Introduce some kind o
På fredag 02. mai 2014 kl. 02:17:58, skrev Craig James mailto:cja...@emolecules.com>>: On Thu, May 1, 2014 at 4:26 AM, Andreas Joseph
Kroghmailto:andr...@visena.com>> wrote: I have a schema
where I have lots of messages and some users who might have read some of them.
When a message is read by a
