column_type override not being honoured for dbt seed

[sdairs]

create a seed with 2 cols

country_code,population
ABW,106766
AFE,677243299
AFG,38928341

Running dbt seed fails due to type INT being inferred, but some values are too large for INT and require BIGINT.

12:56:16  Completed with 1 error and 0 warnings:
12:56:16  
12:56:16  Runtime Error in seed ref__populations (seeds/reference/ref__populations.csv)
12:56:16    AnalysisException: Possible loss of precision for target table 'dbt_demo_reference.ref__populations'.
12:56:16    Expression '3332105361' (type: BIGINT) would need to be cast to INT for column 'population'

Create a schema.yml for the seed to overrise the inferred type to BIGINT

version: 2

seeds:
  - name: ref__populations
    description: Worldwide population data by country
    config:
      column_types:
        population: BIGINT

Use dbt seed --full-refresh to ensure the new schema is used.

12:40:25  Runtime Error in seed ref__populations (seeds/reference/ref__populations.csv)
12:40:25    AnalysisException: Target table 'dbt_demo_reference.ref__populations' is incompatible with source expressions.
12:40:25    Expression ''106766'' (type: STRING) is not compatible with column 'population' (type: BIGINT)
12:40:25 

Fails as the over ride is not honoured.

Sounds similar to dbt-labs/dbt-spark#139 seen on dbt-spark

[tovganesh]

Did some debugging, apparently the table is correctly created, but some issue with the insert:

`create table s3test.ref__populations (country_code string,population BIGINT)

      insert into s3test.ref__populations (country_code, population) values
      (%s,%s),(%s,%s),(%s,%s)
      
      ['ABW', '106766', 'AFE', '677243299', 'AFG', '38928341']

`

[tovganesh]

#18

@sdairs can you verify if the above PR works at your end? This is mostly straight out of spark adapters' solution for similar issue

[sdairs]

Thanks @tovganesh - works perfectly

[sdairs]

Fix has been merged