I have some code that loops through about 1/4 of a million records, it is taking a very long time (I stopped it after 15 minutes and it wasn't close to done).



I took out all the processing in the loop so that all it was doing was the loop with a single seek() and it was still slow. I replaced the seek() with "if 1 =2 then" and it went as expected (about 20 seconds).



I am seek()ing against a single integer field that I have an index on in SQL. Help????



Code:

searchstr = "rent_RPMrecnum = " & "'" & currow("record_number") & "'"

'If Me.RentitemsBO1.Seek(searchstr) Then

‘do nothing

Else

‘still do nothing for testing purposes

End if





Thanks.










Keith Chisarik

Part of the problem is that the volume isn't in SQL exclusively, I have to read a table of roughly 1/4 million records from a Dataflex table via ODBC (seems all I do these days is read old data but thats another story).



I am trying to create a sync method that will mirror the table in SQL for me to program against. So for each of the 250,000 records I need to either add a record or check the fields to see of they have been updated. It is the initial "look for this record" that is taking the time. I think what you are saying here is that I want to create a SQL proc to search for the record, not rely in the BO seek() method since ADO wasnt built to do that... do I have it right?



Thanks everyone for your input and time. For the time being I have abandoned doing the whole table at once, and gone for only updating the records the user calls for at any given time, I will eventually need to be able to "refresh" the entire SQL table against the live Dataflex datasource for reporting though.

Keith Chisarik

Good old DATAFLEX

Are they still around? I used it from 1983 to 1985. It's proprietary data file format (back then) was very solid.

Are they using Btrieve now? Or Pervasive SQL?

..ßen