I am using a combination of tabledefs.delete to remove tables followed by SQL: UPDATE, and DELETE to remove records of that table in 4-5 other related master-tables.

This works ok in principle. But in some cases when there is a need to remove large numbers of tables (and associated records), it is somewhat slow to start, but it even slows down a great deal as more deletions complete.

My only guess is that the database, being large enough(>1.5Gb), becomes more and more fragmented with each delete, and that causes slower operation in general.

Any ideas on how to improve this will be appreciated.

__________________
Tim

If you can adjust your deletion SQL so that it defines more of the records to be deleted, that will run faster than a loop that deletes them 1 at a time.

Where is it that you are using .UPDATE?

Could you illustrate your arrangement with some of the VB that accomplishes this task?

Thanks.
Here is an example, if a table (called symName) is deleted all master tables should be UPDATEd:
    SQL = "UPDATE DataSets INNER JOIN DataSetSymbols AS S ON DataSets.DataID = S.DataID "
    SQL = SQL & " SET DataSets.lastEdit = Now() WHERE (((S.Symbol)='" & symName & "'))"

You are correct in your inference that the deletion is done one a time.
The "mass" delete was added as a giveaway, and it simple loops through individual deletes.
Somehow it has become a critical tool in data migration, so its crude nature is showing.
So it is not optimized at all. For one, it connects and disconnects to the same DB for each delete.
But still that does not explain the incremental slow-down with the number of symbols.

Any hints on this will be appreciated.

Tim