@kapa.ai Hello, I’m trying to build the Elementary models using:

dbt run --select "elementary.*"

but I consistently run into the following error while setting up Elementary UI/reporting:

File ".../dbt/adapters/databricks/impl.py", line 484, in get_columns_in_relation
    use_legacy_logic = relation.is_hive_metastore() or self.compare_dbr_version(16, 2) < 0
AttributeError: 'dict' object has no attribute 'is_hive_metastore'

Environment details: • dbt version: 1.10.6 • elementary version: 0.16.1 I also tried upgrading Elementary, but the error persists. Has anyone encountered this before, or is this a known compatibility issue with dbt-databricks / DBR versions?

The error you're seeing:

AttributeError: 'dict' object has no attribute 'is_hive_metastore'

is a compatibility issue between Elementary and a newer version of

dbt-databricks

. This is consistent with a broader pattern of breaking changes introduced in

dbt-databricks

versions around

1.10.x

and above, where Elementary's internal macros receive a

dict

object instead of a proper relation object, causing attribute errors. [Databricks report errors] A very similar class of errors (e.g.,

'dict object' has no attribute 'config'

,

'Macro' object has no attribute 'config'

) has been reported by multiple users running Elementary with newer

dbt-databricks

versions (

1.10.0

,

1.10.2

,

1.10.3

). [Databricks report errors] Known Workarounds 1. Downgrade

dbt-databricks

: The most consistently reported fix is to pin

dbt-databricks

to a version below

1.10.2

. For example, users confirmed that downgrading to

dbt-databricks==1.9.7

resolved the issue. [Databricks report errors] 2. Use separate virtual environments: One workaround is to run the Elementary report generation in a separate environment with a compatible (older)

dbt-databricks

version, while your main workflows use the newer version. [Databricks report errors] Important Note on Your Elementary Version You mentioned you're on Elementary

0.16.1

with

dbt

1.10.6

. Note that

dbt-core 1.10.x

is a very recent version, and Elementary

0.16.1

is quite old. There may be additional incompatibilities beyond just the

dbt-databricks

adapter version. The Elementary team has noted that CLI and dbt package versions must be kept in sync — for example, CLI

0.16.2+

requires dbt package

0.16.3+

. [version upgrade issue] Recommended steps: • Pin

dbt-databricks

to

1.9.7

(or below

1.10.2

) as a short-term fix. • Ensure your Elementary CLI and dbt package versions match. • Check the Elementary release notes for compatibility with your specific

dbt-core

version before upgrading. The knowledge sources do not contain a confirmed permanent fix for this issue with

dbt-databricks >= 1.10.x

, and it appears the Elementary team was still investigating it as of the last available information. [Databricks report errors] 👍 0 button 👎 0 button ℹ️ Tip: Mention @elementary.ai in the thread for follow-ups or DM the app for 1-to-1 help.

@kapa.ai is there any other work around for this, i tried changing versions but then different errors come up.

Based on the knowledge sources, there is one additional workaround mentioned beyond version pinning: Use separate virtual environments: Run the Elementary report generation in a dedicated Python environment with a compatible (older)

dbt-databricks

version, while your main dbt workflows continue using the newer version. This avoids forcing a downgrade across your entire project. [Databricks report errors] One community member described it this way: > "I made the process that generates the elementary report run with an appropriate

dbt-databricks

while my workflows run with a newer version." Beyond that, the knowledge sources do not document any other confirmed workarounds for this specific

AttributeError: 'dict' object has no attribute 'is_hive_metastore'

issue on Databricks. The underlying bug was tracked in GitHub issue #1931, but as of the last available information, a permanent fix had not yet been released, and users were still waiting on a resolution from the Elementary team. [Databricks report errors] I'd recommend: 1. Opening or commenting on [GitHub issue #1931](https://