The main reason why I’m still using the legacy reports as opposed to the new report format is because one can filter the “today” column of the report very easily in the old format. The column is recognized as a date column in excel and you can filter by month, year etc.
In the new format, the date field is stored as text and there’s no way to quickly filter a full month or year.
Request you to please fix this issue, as the downloading of the new report is more straightforward and I’d really like to start using it.
Dear @Josh,
Great, thanks!
I would suggest, please, to do the same for KoBo export of numeric variables from a form (as it is done already for KoBo _internal vars, like _index).
Also _varname (leading underline) is invalid for SPSS.
Kind regards
Hi @wroos, I am currently working on the data-type export issue at the moment. Thanks for noting the issue with leading underscores and SPSS. In your experience, is it preferable to strip the leading underscore (or similar solution to keep the variable) or discard variables with leading underscores?
Hello,
Keep them ALL, please.
As current workaround: We just add/edit in Excel a prefix like “SYS …” before, e.g. SYS_id. (And also replace some __ inside with _).
For the dates, the most critical for transfer is the datetime type. Different to date, there is no simple edit option with Excel.
Another issue is the “/”, e.g. in the multiple choices. (Invalid for SPSS).
All future changes in the export file should correspond to the SPSS Label export, please.
I doubt, if we want the following (see https://github.com/kobotoolbox/kpi/issues/1266):
“Select_One/Select_Many: If the response is a number, it should be stored as a number and not as a text string”.
There are Pro/Cons: If it becomes numeric, SPSS will probably change the MEASUREMENT LEVEL on import to Ordinal or Scale, instead of Nominal now (which is in most cases statistically correct). Also, such format change text to number is easy to do in SPSS.
