Hi Marco,

I agree with you. In a quick way:

1) My post was just to 'remix things up'.
2) In your explanation, you talk about select * from blabla: in a fast reading, we look at the * and not to the from 'blabla'. There's no where clause, I don't assume it, I was just pointing to the Where clause, and trying to let him forget the '*'.
3) I don't think that speed could be a problem fetching the 10000 rows. He is not complaining about that. He is complaining about the delay that exists after a few hours of working with the database.
4) Yes, it's a nonsense to get a 10000 row grid to be open all day. I think you have give a good glue to client/server environment in here.
5) And yes, we have spoke about transactions in the first line, but all of your text was about reducing the rows fetched from the server. I just point out to the need to explore the transactions problem instead of the number of rows.

We are all trying to help. I just think that if he reduces the number of rows is just delaying the problem from 6 to 12 hours. I really think that this is a problem with very long transactions.

:)

Artur Anjos



[Non-text portions of this message have been removed]