[BUG] databricks column-level tags for materialised tables

[timlu98]

Describe the bug

Seems like column-level databricks_tags are working fine for views but when materialising tables, column tags are not applied (table-level tags work fine via config parameter). Could not find this as a known issue.

Screenshots and log output

System information

dbt_version: "2026.1.20+e69724e"
dbt_databricks_version: "1.11.3"
databricks_sql_connector_version: "4.1.3"

Using dbt cloud

[hazvk]

+1 I have also experienced this issue before.

[jamiepgray]

Issue also experienced on

Core: "1.10.11"
databricks adapter: "1.10.12"

switching the table to a view applies tags as expected.

[jamiepgray]

As a follow on, I am seeing alter statements run for table materializations when setting the dbt_project.yml flag.

flags:
  use_materialization_v2: true

@timlu98 and @hazvk are you seeing this with the V2 materialization strategy?

[hazvk]

As a follow on, I am seeing alter statements run for table materializations when setting the dbt_project.yml flag.

flags:
  use_materialization_v2: true

@timlu98 and @hazvk are you seeing this with the V2 materialization strategy?

I think this is expected, some of the ALTERs need to happen (I was having a read of the codebase and noticed some still there). Eg you can't apply column tags on CREATE, they need to happen using ALTER. Maybe that's what you're seeing?

[sd-db]

As a follow on, I am seeing alter statements run for table materializations when setting the dbt_project.yml flag.

@jamiepgray use_materialization_v2: true correctly applies column tags while table creation, while this is missing in v1 flow. The fix should be simple to add. One thing I have a doubt on (and why v2 was created in the first place) is that applying column tags for large tables can be expensive (server side behaviour; not much in our control). I will raise a PR for the fix and followup internally on any concerns regarding perf.

I think this is expected, some of the ALTERs need to happen

@hazvk can you detail more? Applying column tags to a table will always be an alter operation, currently the SQL syntax does not support adding column tags as part of table creation(hope to be fixed soon 🤞). I guess this is what you are referring to...

[hazvk]

Yes sorry if I wasn't clear that's what I meant. Look forward to a fix on V1, it'll be much appreciated and keep us actively using column tags!

…

On Sat, 7 Mar 2026, 3:19 pm Shubham Dhal, ***@***.***> wrote: *sd-db* left a comment (databricks/dbt-databricks#1307) <#1307 (comment)> As a follow on, I am seeing alter statements run for table materializations when setting the dbt_project.yml flag. @jamiepgray <https://github.com/jamiepgray> use_materialization_v2: true correctly applies column tags while table creation, while this is missing in v1 flow. The fix should be simple to add. One thing I have a doubt on (and why v2 was created in the first place) is that applying column tags for large tables can be expensive (server side behaviour; not much in our control). I will raise a PR for the fix and followup internally on any concerns regarding perf. I think this is expected, some of the ALTERs need to happen @hazvk <https://github.com/hazvk> can you detail more? Applying column tags to a table will always be an alter operation, currently the SQL syntax does not support adding column tags as part of table creation(hope to be fixed soon 🤞). I guess this is what you are referring to... — Reply to this email directly, view it on GitHub <#1307 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA5IZY6CHLKVYUOOBW7XMOD4POPM7AVCNFSM6AAAAACSPFV3YOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DAMJVGUYDCNRSGM> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[jamiepgray]

Thanks @sd-db for raising that pull request. Having this work properly in V1 is ideal.

I know it's outside of the scope of this original bug report, but I'd be curious to know more about the application of column tags being expensive on large tables. Any public facing docs we could refer to to better understand?