synapse and mssql bugfixes and improvements

[jorritsandbrink]

Description

This PR

• adds chunk separation handling for the select_union insert values writer type to fix select_union insert values writer type fails with buffered writing #1173
• sets the max_rows_per_insert property to prevent "Some part of your SQL statement is nested too deeply. Rewrite the query or break it up into smaller queries." errors
• adds mssql and synapse to set of tested destinations in test_insert_job_client.py for better test coverage
• removes pipeline dependency in SynapseStagingCopyJob.generate_sql by passing the job_client as argument
• improves PyOdbcMsSqlClient.rollback_transaction such that it won't fail if there is nothing to rollback
• reorganizes logic in some places for better code quality

Related Issues

• Fixes select_union insert values writer type fails with buffered writing #1173

[netlify]

✅ Deploy Preview for dlt-hub-docs canceled.

Name	Link
🔨 Latest commit	ee1910b
🔍 Latest deploy log	https://app.netlify.com/sites/dlt-hub-docs/deploys/661198f3073be300089b340e

[rudolfix]

Please take a look at test_insert_job_client, you should add synapse there:

DEFAULT_SUBSET = ["duckdb", "redshift", "postgres"]

and also mssql (it also uses inserts)
you can then merge your test into it

[jorritsandbrink]

@rudolfix I added mssql and synapse to test_insert_job_client.py and made some changes so the tests pass. Also removed the pipeline dependency to obtain a job_client in SynapseStagingCopyJob.generate_sql as discussed on Slack—this is done in an unsophisticated way that violates from_table_chain and generate_sql interfaces because the proper way would involve substantial code changes, type checking errors are ignored. See #1174 (comment) for an overview of the different changes contained in this PR.

Can you review?

[rudolfix]

LGTM! thanks for improving the tests!!