Hi,
SQL Anywhere database, to be used by a Powerbuilder 11.5 application.
We're at the stage of designing a set of 6 tables, all of which cascade down to the next one, like
Country->Region->District->Organisation->Group->Member
The dilemma that we are facing now is that the last three tables will have about 15.000, 30.000 and around 200.000 rows, for which we need to generate long random IDs. We have never used UUIDs as primary keys so far and we are considering using them as primary keys for these last three tables which will have thousands of rows.
However, having no experience with UUIDs, we are reading up on them now, and there are several articles suggesting that UUIDs are not very efficient, especially if we link up these last three tables with foreign keys as described above: Organisation->Group->Member.
So our question is: with these table sizes, are we likely to run into noticeable efficiency problems with using UUIDs as primary/foreign keys? And if so, what would be the more recommended approach?
Cheers