Support for `distinct_on` argument in MSSQL (for GraphQL operation)

### Is your proposal related to a problem?

The `distinct_on` arg in GraphQL operation is not supported currently for Hasura connected to MS SQL data source . It's only supported for Postgres and BigQuery. The common use case is that developers/adopters of MSSQL can leverage functionality where they can serve unique/distinct records to their end-users, ensuring that they only get each value once and no duplicates are included in the result.

Ref for current `distinct_on` argument (PG) and its use case - https://hasura.io/docs/latest/queries/postgres/distinct-queries/#the-distinct_on-argument

Start by comparing the existing PostgreSQL and BigQuery implementations of the distinct_on GraphQL argument, then inspect how GraphQL operations are handled for the MSSQL data source. Done means MSSQL users can request distinct records through distinct_on without duplicate values.

Written by the indexing model from the issue text.