Simulate Set RowCount = 100 during Data Sync

ccollins

I am generating the table differences by table and the migration sql script by table already. This is helping with optimizing disk space. In order to improve this, I need to synchronize in batches smaller than the entire difference record set.

I have a table that is very large with respect to data, not record count.
I want to update the table in 100 record batches until the table is in sync.

I will be refreshing the comparisionsession object after each batch.

What is the best method of selecting or filtering what is generated during GetMigrationSQL.

A code example would be helpful.

Brian Donahue

Hi,

This is a good question. I don't know a way to change the behaviour of GetMigrationSQL. Executing the individual batches in the ExecutionBlock probably won't help because the batches can encompass updates for an entire table anyway.

Is there something that can be done with a Stream object that could help? You can use the GetFileStream method of the ExecutionBlock to obtaing a FileStream object that will let you stream the ExecutionBlock, but I don't know if you would find that useful or not.

ccollins

I tried the streaming suggestion. However, the temp file generated for the table update is still larger than what we would like to see at one time, (> 600MB).

We are attempting to reduce the total space required to update the data. The RGDC- files take up enough space before we even begin to generate the sql migration temp files.

I was breaking the update down by table, but I believe I run into an unexpected behavior doing this: No disable trigger or disable FKeys. This unexpected behavior has me trying to first figure out how to get my table updates in the correct order before I can go back to trying to limit the size of the .TF temporary temp file.