Incremental merge or insert+update fail if configuring composite unique key

[jaceksan]

Example:

  materialized='incremental',
  unique_key=['pull_request_number', 'repo_id']

Using dbt-vertica version 1.5.0.

[jaceksan]

It generates query such as:

delete from "XXX"."github_input_stage"."pull_requests_extract_json"
where (['pull_request_number', 'repo_id']) in (
  select (['pull_request_number', 'repo_id'])
  from "pull_requests_extract_json__dbt_tmp"
  ....

[barloc]

Hello.
Use it via concat or hash unique_key='hash(load_dt,table_schema,table_name)',. Adapter doesn't support list for this option, only string value.
From documentation: The unique_key config parameter is required for using the merge strategy, the value accepted by this parameter is a single table column.

[jaceksan]

I see.
On the other hand, the list is supported by Snowflake ;-)

[ajay-abrol2]

Working on to support unique_key parameter as a list. So it will always accept list not string value even for single value.

Examples are below:
for single column, configuration should be : unique_key=['columname']
for multiple columns, configuration should be: unique_key=['col1','col2','col3']