Jdbc connector - limit varchars for sink process

Hello,
I facing a performance issues when writing to a MSSQL cluster with jdbc sink connector.
The issue is located in the fact that there is a “string” as part of the PK. (I have tested it without this value and it was super fast)

Doing some query analyses it turned out the connector is writing all strings as nvarchar(4000). This leads in some performance issues. Is there anyway (transform etc) to limit a string?

The customer_id unfortunately is based on a md5hash so I can’t CAST it to integer etc.

A workaround would be to use another field as PK tho. Nevertheless I would like to stick to this PK if possible.

Thanks for your help in advance.

1 Like