👋 I tried to run a pipeline tap-postgres target-snowflake, but I get the error:

Extraction failed              code=1 message=ValueError: Out of range float values are not JSON compliant

This is because in my tap table, I have a double precision column with a "inf" value. I have a few of these columns across a couple of tables, so I'll need to repeat the solve a few times. Has anyone found a workaround for this?

Hey @Simon Schmitke! This is definitely an edge case of serializing data as JSON 😅. I think it's solvable in the singer-sdk via `simplejson`'s `allow_nan`/`ignore_nan` parameters, but I'm curious about what the use cases are for this type of data. Is Snowflake able to load an "inf" value? Would it be ok if tap-postgres made these values null?

Thanks @Edgar Ramírez (Arch.dev) I'll see what I can do.

I'm curious about what the use cases are for this type of data

I'm looking for alternatives to fivetran for our companies production database. We have a few tables that use this logic. Example: Customer is allowed integrations in the amounts of: • 0 • 1 • 2 • 3 • inf So NULL can work and we'll just update some downstream logic in dbt so I think it's fine, but our app works with the

inf

as-is