The PK column (systemstop) should be of type smallint, but it looks like it's being ingested with a .0 tacked on (presumably by pandas/geopandas). None of the records have a null value in that column, so I'm not sure why pandas/geopandas would treat these as decimal values, but it is.
Anyway, the fix is pretty simple. Modify the relevant {dataset_name}_standardized.sql just split it at the decimal and cast the first element of the resulting array to smallint
The PK column (
systemstop) should be of typesmallint, but it looks like it's being ingested with a.0tacked on (presumably by pandas/geopandas). None of the records have a null value in that column, so I'm not sure why pandas/geopandas would treat these as decimal values, but it is.Anyway, the fix is pretty simple. Modify the relevant
{dataset_name}_standardized.sqljust split it at the decimal and cast the first element of the resulting array tosmallintthen run `dbt from that pipeline stage onward.