These are chat archives for frictionlessdata/chat

6th
Sep 2017
Robert Gieseke
@rgieseke
Sep 06 2017 11:20
Congrats on 1.0!
It seems that year and yearmonth are not supported by all tools, e.g. the Python goodtables lib issues a warning
datapackage.json" has a validation error "\'year\' is not one of [\'string\', \'number\', \'integer\', \'boolean\', \'object\', \'null\', \'array\', \'datetime\', \'date\', \'time\', \'duration\', \'geopoint\', \'geojson\', \'gyear\', \'gyearmonth\', \'any\']"',

And with

datapackage.push_datapackage(url, 'pandas')

I get "ValueError: Cannot cast '2017-01' for <yearmonth>"

Which repo would I create an issue in?

Paul Walsh
@pwalsh
Sep 06 2017 11:38

hi @rgieseke

The core libraries (datapackage and tableschema) have been updated to V1, but the higher-level libraries are in process now (e.g.: goodtables, the various storage drivers like pandas, sql and bigquery). Sorry for the confusion!

Robert Gieseke
@rgieseke
Sep 06 2017 11:43
Thanks for the quick reply! So is there a underlying shared library that both goodtables and datapackage use?
Paul Walsh
@pwalsh
Sep 06 2017 11:51
@rgieseke you have problem with datapackage too? from examples above i assumed only with the pandas integration with the DP library
(and goodtables)
Robert Gieseke
@rgieseke
Sep 06 2017 11:57
@pwalsh Sorry, only tried the pandas integration!
roll
@roll
Sep 06 2017 12:13
@rgieseke We're updating goodtables and sql/pandas/etc integrations right now. There is a general issue for it frictionlessdata/implementations#22 (in your pandas case).
Robert Gieseke
@rgieseke
Sep 06 2017 12:19
@roll @pwalsh Thanks both! Subscribed to the issue ...
roll
@roll
Sep 06 2017 12:23
@rgieseke you're welcome)
Robert Gieseke
@rgieseke
Sep 06 2017 12:32
btw goodtables-py is awesome! very useful and already finding problems with my csvs ...
Paul Walsh
@pwalsh
Sep 06 2017 12:33
Thanks @rgieseke ! we think so too. You can also check out https://goodtables.io/ for data validation as a service ;)