I read on stackoverflow that locking can be a problem when using process-keyed tables, and I was wondering if you had any suggestions on how to minimize or avoid unnecessary contention (e.g. table locks) when using these types of tables. Here's some background:
I'm developing a batch process that inserts the rows resulting from a stored procedure call into a table. I have no control over the stored procedure. The table is keyed by process id (@@spid) because it needs to be read by several stored procedures that follow and should not conflict with other batches running at the same time. This shouldn't be too big a problem because it is a low volume server, e.g. not too many of these batches will be running simultaneously. Still, I wouldn't want to create table locks if it can be avoided as I imagine blocking may cause unexpected failures or delays.