-
Notifications
You must be signed in to change notification settings - Fork 22
Missing columns + incorrect date value (Good Tables report) #16
Comments
I've figured out the missing value things. The non-castable value is going to be a problem—we do have to deal with dates BCE. |
Just as a status update, your review has triggered a much-needed and comprehensive review of our export process. Expect new data very soon, now with additional oversight! |
@workergnome that's excellent! Let me know where I can help! |
Hiya, Dan. New data's up, if you want to try your magic data tools on it. |
✨
Based on the following, it looks like the multi-value cell issue (frictionlessdata/datapackage#381) needs to be addressed. This causes the many Some rows don't have a
|
So interesting. So, date_acquired (which is, as you might guess, the date that we acquired the work) is, in fact, a required field. 650 works in our collection don't have one of those, which means that we're now going back and fixing those records. |
I forked this repository and slightly edited the
datapackage.json
to remove the schema references which now work differently (http://specs.frictionlessdata.io/tabular-data-resource/). Unfortunately, this involves literally copying the schema across two different resources. This is necessary to use the tools until they catch up to v1 of the specs and support JSON Pointers.After that, I ran
goodtables datapackage datapackage.json
(https://github.com/frictionlessdata/goodtables-py) which will test the first 1000 rows. It looks like the[missing-value]
errors result from having too few columns (commas) for these rows. It looks like several columns look something like below:as opposed to:
Finally, on row 991, there is an incorrectly formatted date:
-0001-01-03
.There may be more errors like this. This is only for the first 1000 rows.
The text was updated successfully, but these errors were encountered: