We have some proprietary software that processes print files (AFP) and within the scripting for a process it makes a call to a SQL server via ODBC once per page in the file. This might mean something like 50,000 calls per file which seems quite intensive. The lookup data is simple, approx 12,000 rows across 4 columns, all varchar(5) and in the code we call a stored procedure via ODBC. This might be a limitation of the software rather than ODBC itself but we cannot process 2 print files at once unless we create copies of the ODBC with different names and point different files at different connection names even though they point at the same SQL SP.
If we move to the datasource being a local csv file instead of SQL, would the csv file get locked by the first call regardless of the ODBC name? It looks like we might need to continue the multiple names of ODBC connectors if it is a limitation of the post composition software but would that then cause issues using csv instead of SQL as the incoming connection to the source csv file could be coming multiple times from either the same ODBC or "different" ODBC connections?
