A

9 Messages

 • 

550 Points

Wednesday, July 10th, 2024 1:52 PM

dbt incremental models with the Snowflake technical lineage connector

We have recently deployed the snowflake technical lineage capability via Edge.  We're using a JDBC connection in SQL-API mode.  We have a number of teams using dbt for transformation and we've noticed that technical lineage doesn't always work for incremental models.  I'm not sure but this mainly seems to affect dbt versions > 1.1.  I can't see any obvious differences in the SQL generated so I wondered if anyone else had come across this issue?

I'm aware of the specific dbt connector but this will be really onerous for us to deploy as we have quite a decentralised org structure so I'm hoping to get the snowflake lineage connector working in as many use cases as possible.

9 Messages

 • 

550 Points

2 months ago

In case anyone else has this issue I tried a couple of things which seemed to make an improvement:

  1. Adding this under "Source Configuration" on the Edge capability (as some dbt incremental models use temp tables to stage data under the hood)
    {
         "analyzeTemporaryTables": true
    }
  2. Following these instructions to update the configuration to cater for dbt dropping and replacing objects.  Although this shouldn't be related, I only got lineage for incremental models when I did both of these steps.
    Snowflake SQL-API lineage missing due to frequent table replacement – Support Portal (collibra.com)

The other way to achieve this seems to be to use the SqlDirectory option and submit relevant sql statements directly.

Loading...