Open richplane opened 9 years ago
Confirmed this issue. A letter 'T' in cell, as might be used to indicate 'true' or 'false' - is also interpreted as a timestamp datatype.
We still experience this problem on version 2.7.3. For example datapusher changes the data on CSV file from 11.5.1998 to timestamp 1998-11-05T00:00:00 Is this fixed on newer versions?
Our client is uploading CSV data to CKAN which includes a column containing a day of the week.
Datapusher (via messytables) appears to decide that this is a date column and converts it to a timestamp. "Wednesday" appears to be interpreted as "This Wednesday" and is replaced by a timestamp in the form 2015-06-03T00:00:00 - which is useless to anyone reading the data, and frequently misleading.
Since the messytables code seemed to say that DateType includes format detection, I tried altering the configuration to use that instead of the DateUtilType that Datapusher uses normally, by adding:
into datapusher_settings.py. But it made no difference.
Is there any way of ensuring that dates get recognised as dates (and don't get given a spurious 00:00:00 timestamp), and days get left as days (strings)?