What about creating dummy tables in AREV 3.12 to port the Rev/G data into first and then attach these in OI/ARev32 and copy the data into their final destination? You don't even have to create dictionaries for these dummy tables. They are simply there to be a staging ground in lieu of having to export all the data to CSV files (or whatever).
That's definitely doable, but it would take a Write for every record. I'm creating a DOS file for each RevG file, and OSBwriting the data out in 60K chunks. On the Arev32 side I can do the same, OSBreading in chunks.
Would writing and reading to/from an Arev table really be faster?
YMMV, of course, but I would certainly try it first to see how well it might work. Definitely presize the dummy tables to be larger than you need and then sizelock them to prevent dynamic sizing. That will give you optimum speed in the copying process.
Comments
Would writing and reading to/from an Arev table really be faster?