I had to create an application once that had potential new records coming in for the same table from multiple distributed databases, that merged together with manual replication.
I started with Integer ID's, but quickly realized I had to use UniqueIdentifiers to avoid conflicting ID values.
So I got into the habit of ALWAYS using UID's in tables vs. integer ID's. And later I met an expert SQL programmer who said I was doing the right thing for larger systems, and that it was a good habit to develop.
But I still see a lot of other SQL database developers still using integer ID's, even by expert, long-time SQL developers.
So what's the "best practice"?
When is it preferable to use integer based ID's vs UID's?
2007-07-06
01:59:28
·
4 answers
·
asked by
Anonymous
in
Programming & Design