>I have a form that inserts a couple of million records into a table from a text file. This process seems to get slower and slower the longer it runs.
>
>I would like to delete the indexes on this table before reading in the data. Hopefully this will speed thinks up.
>
>I found the delete tag command that actually deletes the cdx file. Is there a way to recreate these indexes programmatically when everything is done. I looked at the index command, but I don't think this will work because I need on of the indexes to be primary.
>
>Is there another way to do this?
>
>Thanks
>
>Pat Murphy

Pat,
If you do delete the indexes, it may speed up the import (somewhat) but the problem will be recreating all the indexes on millions of records. You will have to go through the table once for each index and that will probably take longer than waiting for the process to update all the indexes initially.

If you don't need all the indexes, delete all but the primary key index (updating 1 index during an import takes very little time). Then create indexes as a user requires them. Because there are a lot of records, there will be a significant delay when a user requests it. But maybe you (and they) can live with that.

Just my opinion.