I've got a file that contains several external data source tables from a JD Edwards EnterpriseOne (Oracle 11g) database.  If I understand correctly, FileMaker doesn't just execute a query via the ODBC connection, but rather downloads each connected table in its entirety when the file is opened.  Is that correct?  The problem with this is that the JDE tables are very, very large!  Each one can have a large number of fields/columns, and thousands of records!  This means that a ton of superfluous data must be transferred over the LAN.  (I only need a couple fields, and only from records that would match a WHERE clause.)  This makes the FileMaker file very slow to load, and very slow to change layouts.

There must be a better way.  Anyone have any suggestions?

Where did you hear that?

The most efficient way to do what you want is to set up a view on the SQL back-end with just the fields you need.  And connect to that instead of to the real table.

Ha ha!  Judging by your reaction, I'm way off base.  I thought that's what I had learned on this site, a couple years ago.  Maybe it was somewhere else.  It's been a while since I've done much with the file in question.

So then, FileMaker doesn't transfer or read the entire table when the file is opened?  Then what would cause it to have a delay of several minutes before anything is displayed?

Thanks for helping to further my (limited) understanding of the technology!