Hi all. I'm having an issue when using the dbt ingestion CLI to create glossary terms from the meta_mappings. (using version

v0.9.2

of the CLI and of DataHub). • When created, I can see the Glossary Term on the DBT Dataset in DataHub (screenshot) which has been automatically assigned using

meta_mapping

. • If I click on the glossary term, it says it exists and I can see all the related entities (screenshot) • However, if I go to the glossary, the glossary term isn't displayed (screenshot). • If while on the phantom glossary term's menu where I can see the entities etc, if I click on the three dots and try to move it into a term group, it says "Unkown Error Occured" (screenshot). I've noticed that I can actually type absolutely anything into the URL urn (

/glossaryTerm/urn:li:glossaryTerm:<ANYTHING>/Related%20Entities?is_lineage_mode=false

). I can type any string into where I've put

<ANYTHING>

and it'll give me a glossary view of that term. However when they're created in the UI, they are given a long uuid which you can see in the URL. • Terms created in the UI persist in the Glossary menu, and can be moved into groups • Terms created via the datahub ingestion CLI (the API) can't do either of the above • Creating them in the UI first, then syncing the terms up via the ingestion CLI doesn't link the term you created in the UI to the term you've assigned to your datasets via

meta_mapping

, because they seem to have different

urn:li<BLABLA>

values. The UI is a uuid, and the dbt ingestion cli one is a friendly name. Any ideas? Edit: I believe this issue is related but not a duplicate of this: https://datahubspace.slack.com/archives/C029A3M079U/p1666343681646089 (cc @bulky-soccer-26729 @gifted-bird-57147)

cc @gray-shoe-75895 -- This problem of "reference-only" ingestion has come up before. We should ideally be minting the glossary term entities prior to submitting the FK references to them here

Thanks @big-carpet-38439, if there's a bug issue open for this in GitHub or a PR, please let me know and I'll do my best to contribute. My Java is awful but if it's part of the python codebase I should be able to help.

For now the easiest thing to do is to pre-create the terms in the Glossary section of DataHub, and then map to those using their URN in Meta Mappings

The meta_mapping should go in your datahub ingestion config, and the meta.company field should be in your dbt schema.yml

How should configure meta.company field in our dbt schema.yml? can you please post an example?

Here’s an example config https://github.com/datahub-project/datahub/blob/master/metadata-ingestion/tests/integration/dbt/test_dbt.py#L170-L186, and the corresponding dbt schema yaml https://github.com/hsheth2/sample-dbt/blob/main/models/base.yml