|Perl: the Markov chain saw|
If you're worried about unusual table namesWell, no - that's not my problem.
The problem is that I will not know the names of the tables until run-time.
Let me give another (once again, contrived) example, that might explain it a bit better.
Imagine that I have two sets of tables, one a set of transactional records (eg, sales records), and the other a set of consolidated tables for reporting. So I have a Sales table with attributes for Date/Time, ItemID & Price.
Then I have a set of tables that look like Sales201003, Sales201004, Sales201005, etc - a separate table for each year/month combination. (This is not the way I would design it - but this is how it is and so I need to deal with it).
The SalesYYYYMM tables have one row per day, per item, along with attributes for the total number sold, and total revenue.
So I have code that selects records from the Sales table, and then inserts summarised data into the appropriate SalesYYYYMM tables. The code I would _like_ to be able to write would look something like this:
But the above won't work because I can't use a placeholder for a table name.
I can work around it easily enough by creating a separate statement handle for each YYYYMM combination.
I'm just not sure what the best approach to that would be.
Oh, and please ignore the fact that it's horrible data base design, and that calculated data is being written back into the data base. As I said, it's a contrived example that helps to describe my problem.