Implement 'mostly' in expression tests

[clausherther]

We should support the mostly parameter in expression tests as a float that controls what % of rows need to satisfy the given expression test. Default is 100%.

[seanpwlms]

Apologies if this is addressed elsewhere, but this looks like the closest issue. For the expect_grouped_row_values_to_have_recent_data test, I have one group that will not have recent data for valid reasons. Currently the test fails, saying Got 1 result, configured to fail if != 0. It would be great if I could specify fail if result > n.

[clausherther]

Hi @seanpwlms ! I think this might actually be supported in the new dbt test configs:
https://docs.getdbt.com/reference/test-configs
I haven't played around with this, but I think the relevant config is error_if https://docs.getdbt.com/reference/resource-configs/severity
Would love to hear if this is works for you and how you ended up using it!

[seanpwlms]

@clausherther Thank you! Of course the answer is to read the docs. Adding error_if converts the severity to a "warn", which you can make pass by adding a warn_if setting.

[rfortmann-ewolff]

I tried using the severity and the warn_if arguments but dbt encounter an error:

          - name: name_tracker
            tests:
              - dbt_expectations.expect_column_values_to_be_in_set:
                  value_set: ["ssc", "sp2", "cf", "ssc1"]
                  +severity: warn

Running with dbt=0.20.0
Encountered an error:
Compilation Error
  Invalid test config given in models/snowplow/sources_in_snowplow.yml:
        test definition dictionary must have exactly one key, got [('not_null', None), ('+severity', 'warn')] instead (2 keys)

and

          - name: name_tracker
            tests:
              - dbt_expectations.expect_column_values_to_be_in_set:
                  value_set: ["ssc", "sp2", "cf", "ssc1"]
                  +warn_if: "!=0"

Running with dbt=0.20.0
Encountered an error:
Compilation Error in test dbt_expectations_source_expect_column_values_to_be_in_set_snowplow_events__0__name_tracker__ssc__sp2__cf__ssc1 (models/snowplow/sources_in_snowplow.yml)
  macro 'dbt_macro__test_expect_column_values_to_be_in_set' takes no keyword argument '+warn_if'

Any idea what the problem could be ?

[rfortmann-ewolff]

I could solve my own problem.

          - name: name_tracker
            tests:
              - dbt_expectations.expect_column_values_to_be_in_set:
                  value_set: ["ssc", "sp2", "cf", "ssc1"]
                  severity: warn

I hope it might help others.

[clausherther]

@rfortmann-ewolff glad you solved it! I think the + syntax in dbt is for project config settings only, but not sure.