On to another issue, I am running a basic ELT flow on tap-salesforce and target-postgres but the following error keeps popping up: tap-salesforce | Exception: Error syncing User: [Errno 32] Broken pipe tap-salesforce | Exception ignored in: <_io.TextIOWrapper name='<stdout>' mode='w' encoding='UTF-8'> tap-salesforce | BrokenPipeError: [Errno 32] Broken pipe meltano     | Extraction failed (120): BrokenPipeError: [Errno 32] Broken pipe meltano     | Loading failed (1): ValueError: Unsupported column type anyOf: LastLoginDate I researched on my own but cannot figure this out.

@timothy_suh The error messages indicates that in the

SCHEMA

output by the tap, the

LastLoginDate

property is defined as having either type A or type B (

anyOf: [a, b]

) which is not supported by target-postgres since a database column can only have one type: https://github.com/meltano/target-postgres/blob/master/target_postgres/db_sync.py#L81 Can you find out what the type of

LastLoginDate

looks like by running

meltano elt

in debug mode (

meltano --log-level=debug elt ...

) and sharing the offending

SCHEMA

message? Depending on what the

anyOf

type looks like exactly, we may be able to override it with a single type in the catalog file using the

schema

extra: https://meltano.com/docs/plugins.html#schema-extra

schema shows "LastLoginDate": {"anyOf": [{"type": "string", "format": "date-time"}, {"type": ["string", "null"]}]}

All right,

tap-postgres

is tripping over the fact that the second option repeats the

string

type. This would've been fine:

"LastLoginDate": {"anyOf": [{"type": "string", "format": "date-time"}, {"type": "null"}]}

You can tell Meltano to override the schema for the LastLoginData property in the catalog file it passes to

tap-salesforce

(which it hopefully uses to generate its

SCHEMA

messages), like this (with the correct entity/stream ID):

meltano config tap-salesforce set _schema <stream ID> LastLoginDate anyOf '[{"type": "string", "format": "date-time"}, {"type": "null"}]'

I have to do this for every date field - oh boy!

If all the date fields end in

Date

, you should be able to use

*

as the entity identifier, and

*Date

as the attribute identifier, to catch all properties at once!

Trying it now - beautiful solution!

Awesome!