Sorry, that was the statement that caused the error. I was creating a view that exists in the MSSQL land. It actually joins a few tables. I can put a create statement for all the tables used in and then create the view and re send the txt file with those. I am reloading the LINUX from scratch at the moment, but as soon as I get back up (be tomorrow probably as it takes over night to load the data from the MSSQL server) I will email with all the pertinent information.
I am re-loading to hopefully get rid of the pg_user error I was getting (I went to su postgres and created my data base that way after creating a postgres user as root). My friend said to not create any users just start the data base up (Fedora core 3) and use pgadmin to create the database.
There are probably RPMs for 8.0 available by now. Might be worth going to that from the start, rather than upgrading later.
I was following a how to convert I got off the archives, so I must of messed something up.
Again thank you for the information. If it is non TOAST (sorry not sure what that means; I am guessing like not part of a text field data) field sizes adding up to more the 8k is there some way to produce the data set, or is this a limit of Postgres in general. If I can not have all the data needed in a recordset I might have to re-think using postgres is this a limit of mysql also? I hate to think I have to consider staying on MSSQL as it is not in our budget.
Well, that's a lot of non-text columns to breach 8kB. It is a definite limit, but you shouldn't see it until you have hundreds of columns. If you can post the table definitions along with the view definition, that should let people see if they can reproduce the problem.
-- Richard Huxton Archonet Ltd
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match
