The first variant is faster while the second one requires less
maintenance if you rename your documents a lot. Both will do the job.
I'd use the first, de-normalised model unless you normally have a lot
of revisions per document, modify the common document meta-data quite
often, and the meta-data
Ok, I'll go with the first model. Thanks for the reply!
On Jan 13, 12:53 am, Alexander Kojevnikov
wrote:
> The first variant is faster while the second one requires less
> maintenance if you rename your documents a lot. Both will do the job.
>
> I'd use the first, de-normalised model unless you
