tsql

Other adapters (e.g. dbt-spark) have adopted a single-source-of-truth approach to documentation, prefering to document setup and configuration information only on the docs.getdbt website, rather than duplicating it on the docs page and the adapter repo's readme.

I think we should do the same.

Using statistics, see if the max value in a column implies that it would be better as a tiny/smallint. Excluding PKs

original dbt Slack thread in #dbt-synapse

Think I found a bug in the 0.19.2 version of the dbt-synapse.
When making a snapshot with a new detected column, it will try this query:
alter table <db>.<schema>.<table> add COLUMN <name> <datatype>;
This should be:
`alter table ..

add COLUMN