pandas.DataFrame.to_gbq missing method of writting

[OmerJog]

BigQuery support 3 different types of load methods:
WRITE_EMPTY, WRITE_TRUNCATE, WRITE_APPEND
https://cloud.google.com/bigquery/docs/reference/auditlogs/rest/Shared.Types/WriteDisposition

this parameter needs to be controlled by loading job.

[TomAugspurger]

Does pandas-gbq support this?

[OmerJog]

@TomAugspurger yes
https://pandas-gbq.readthedocs.io/en/latest/writing.html#writing-to-an-existing-table
"Use the if_exists argument to dictate whether to 'fail', 'replace' or 'append' if the destination table already exists. The default value is 'fail'."

[TomAugspurger]

So we just need to add a keyword and pass it through? We'll need to take some care to give a nice error message if the user has too old a version. Are you interested in working on this?

…

On Thu, Nov 14, 2019 at 1:40 AM OmerJog ***@***.***> wrote: @TomAugspurger <https://github.com/TomAugspurger> yes https://pandas-gbq.readthedocs.io/en/latest/writing.html#writing-to-an-existing-table "Use the if_exists argument to dictate whether to 'fail', 'replace' or 'append' if the destination table already exists. The default value is 'fail'." — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#29598?email_source=notifications&email_token=AAKAOIWFPLFNT25XTNPO2TTQTT6HRA5CNFSM4JM4JO6KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEEA4TRQ#issuecomment-553765318>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAKAOISDN4JBQRBTOIKZMKDQTT6HRANCNFSM4JM4JO6A> .

[OmerJog]

@TomAugspurger yes. It's just exposing this parameter in the to_gbq function.

I currently don't have the time to do work on this. Maybe i'll be able to in a few months so if someone want to pick it up - be my guest.

[charlesdong1991]

emm, but seems pandas already supports if_exists? or do you mean this argument is not functioning correctly?

https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.to_gbq.html

@OmerJog

Seems a Needs Tests label is more appropriate for this issue since there seems no tests to this case. @jbrockmendel