D
dbeavon
Guest
rkumar, my preference would be that the SQL92 DDL statements would enforce the same validation rules as the corresponding DF. Unfortunately many of our datadictionary operations in Progress are still accomplished via the old-school DF files. (I've found that schema-dumps/reloads rarely work properly with the SQL92 schema.) In short, it seems unwise that the SQL92 statements allow users to subvert the normal validation rules that have always existed in the DF (presumably keywords have previously been restricted for a reason). James, doesn't ANT/PCT rely on the same internal components as dump_df and load_df? I was googling a bit and discovered the ADE development tools code, and my understanding is that it is the basis for dump_df and load_df. And it is probably the basis for whatever is happening deep within ANT/PCT. I would bet that if you used my SQL92 ddl, to include the "brn_bank" table in your schema, then your ANT/PCT would probably choke and die on it too. Basically that one column in the table will prevent an entire empty database from being created via intermediate DF.
Continue reading...
Continue reading...