[CDCSDK] Before image should use it's corresponding schema for row decode

[sureshdash2022-yb]

Jira Link: DB-4610

[sureshdash2022-yb]

As per the discussion with the team, without packed row enabled, for the before image CDC will use the current running schema version to decode the records. which is right now existing current behaviour.

For the packed row enabled with the before image, we will create a separate issue, to pick the corresponding schema version of the before image.