Direct data copy with PG_DSN automatically preserves table case

[necouchman]

I was trying to use the PG_DSN to load data directly from Oracle. I've already loaded up the tables, views, and sequences, and I left the PRESERVE_CASE directive with its default of 0, the result being that the table and view names are converted to lower-case. Unfortunately, when ora2pg goes to do the data load with either the INSERT or COPY type, the case seems to be preserved, and I get several errors like this:

WARNING: destination table SCHEMA.TABLE doesn't exists, aborting data export for this table.

(There are 853 tables in this particular database, and they all fail with this error.)

I looked through the documentation and did not see any indication that this is to be expected, so apologies if I just missed that somewhere in the man page or on the web page. Commenting out PG_DSN and doing the load to a file seems to work properly - the tables are lower-case - but the database I'm working with is in the neighborhood of 500GB, so I was hoping to avoid having to have an intermediate landing spot for the data.

Let me know if I'm just missing something or need to do something differently.

[darold]

It look that you have created the postgresql tables in a different schema. What is the value of EXPORT_SCHEMA and if it is enabled does SCHEMA is equal to PG_SCHEMA?

[necouchman]

It look that you have created the postgresql tables in a different schema. What is the value of EXPORT_SCHEMA and if it is enabled does SCHEMA is equal to PG_SCHEMA?

Thanks for the reply, @darold! The only sense in which this happened is that the schema names were converted to lower-case (SCHEMA -> schema) during that part of the migration - other than case, the schemas were migrated over as-is.

EXPORT_SCHEMA is set to 1, and I did not set a specific SCHEMA, because the Oracle DB has 2-3 schemas that I'd like to migrate over, preferably at the same time.