udpated sql for table comments

[gregboyer]

I still suck at git. Gotta git add before I git commit! This has additional comments for all SBA related tables and columns

[VincentLa14]

Ah! Yes, using pd.read_sql is not really the right way to execute general SQL statements. The point of pd.read_sql is to return a DataFrame object, so generally is used with select SQL statements. To execute SQL in sqlalchemy.

To execute SQL in sqlalchemy you can use connection.execute. See the docs for an example: http://docs.sqlalchemy.org/en/latest/core/connections.html.

In any case it's better that you added the comment stuff in the pipeline code. That way everything touching the database is in one place (and it's easy to recreate the state of the database at any point in time.

[VincentLa14]

Would you mind removing these COMMENT ON statements now that you added the code in pipeline_runner?

[VincentLa14]

Would you mind making pipeline/pipeline_tasks/documentation/table_comments into a sql file? You can just rename it to table_comments.sql. This way we can just execute the SQL commands inside the file.

Also do you mind adding the table_comments.sql file to run within https://github.com/sfbrigade/datasci-sba/blob/master/pipeline/pipeline_runner.py.

You'll likely just want to add it as another element in the list: https://github.com/sfbrigade/datasci-sba/blob/master/pipeline/pipeline_runner.py#L100

[VincentLa14]

hmm might need to be careful with merge conflicts now that @zlatankr has renamed pipeline/pipeline_tasks/documentation/table_comments