I'm embarking on a new project, converting a 30-year old embedded database to Postgres using DF-19. I've tried a Postgres conversion using the Order Entry database and that went ok but have run into a few gotchas with the project database.

The first thing is that in an embedded SQL statement it seems like Postgres requires double quotes around anything that isn't lower case. i.e. if column and table names are in mixed case, the statement: select Name, Address from MyTable would have to be select "Name", "Address" from "MyTable". The only way around this insanity would seem to be to lowercase all of the table and column names in filelist, as well as the file.field names in the .fd and .tag files. Has anyone else encountered this?

Secondly I'm considering converting this database to standard tables but when I did a trial conversion I ran into all kinds of problems with primary keys. I haven't had time to dig into this much more but am guessing the Recnums were the primary keys in most tables. Not that they were used for relationships, they were not. So my choices would possibly be to either leave the Recnums or don't use the automated primary key creation and do it myself. Considerations?

And last, what driver is being used? Is ODBC providing satisfactory speed?