[FEATURE] catalog generation should include rowcounts of models where possible

[dataders]

source: community Slack thread

If the data platform provides rowcounts as a column in the resultset of information_schema.tables, dbt adapters are able to fetch this information via dbt docs generate and store it in the catalog.json, so that it is available in the OSS dbt docs as well as dbt Cloud's Explore product.

one example is dbt-snowflake, which uses the snowflake__get_catalog_tables_sql() macro to fetch Row Count.

from @prdpsvs:

Unfortunately, [Fabric's information_schema] does not [contain rowcount info]. There are multiple ways to retrieve data from sys views/objects but none of them are 100%.
Two approaches can work but both are workarounds and not worth the effort but its upto developers.

• Create stats and query row count using DBCC SHOW statistics
• Use select count(1)

Perhaps one of these workarounds could be shipped in the adapter, but @prdpsvs would know best if doing so would:

• not result in a noticible performance hit for users, and
• be worthwhile if Fabric SQL will soon have rowcount in the info_schema

[prdpsvs]

Unfortunately, this is not currently possible without running dynamic SQL. Depending on number of objects, this could take a while. Warehouse needs to support this with one of the system views. Until then, can't do much