Import with Update

[fmsavey]

Has anyone else noticed the crappy and sometimes fickle performance you get when you do an import with update. I have a solution whereby I need to import records from an ODBC source and the import must update existing records (one match field) in the database. These imports are in the order of 30,000 plus records and the "Records Remaing" indicator counts down one record at a time. At other times when importing the exact same records the "Records Remaining" indicator counts down by hundreds sometimes thousands of records at a time. This is so frustrating. Anyone have any ideas as to why this is happening.

Edited August 10, 2005 by Guest

[fmsavey]

I just realized, after some experimentation, that when I exit and then reopen the database the import goes very fast. If I do the import a second time without closing and reopening the database it takes forever to complete. What is going on here?

[Reed]

The only time I've seen really bad performance was when I was importing while logged in under an account that didn't have view privileges for the fields to be updated. I thought that running the update script with full access would take care of this, but nothing sped things up except giving the account view access to the records.

I've never had to import 30,000 records, but I update ~6000 records every hour using a script scheduling plugin. It seems that the script takes the same amount of time to finish whether it's the first time or the hundredth...

(running the imports on a mac 10.3.9, fmp 7v3, from SQL server 2000 running on windows 2000 server, using the ODBC driver from actual technologies)