Here is a neat trick I just came up with when creating Scribe integration for a customer…
I’m creating a integration where my source is query publisher that is calling T-SQL stored procedure. The destination is some web service, but the web service owner wasn’t very pleased when the integration ran for the first time and web service got called many times and with Scribe’s 8 message processors, more that one simultaneous call was made.
Here is the trick…
I added dummy field with constant value to to query publisher’s query and then specified that I want to group by that column in query publisher settings. But with just a bit more effort I can also specify maximum group size and I speed up integration tremendously! Just use row_number() function and divide the number to group rows.
In the example below, I group 25 rows in one message that gets send to scribe queue.
Here is where you specify grouping. Create grouping node, and drag all fields except group number to <RepeatingElement>.
Check queue for messages and you are done. Beware that the schema of XML message will change and you will have to correct/recreate the DTS mappings.