Odkmeta input csv format

[Matthew_White]

Hi @Naod!

@Naod

I perfectly execute the sample prepared by Matthew that contains date but will get an error (datetime variable SubmissionDate could not be converted using the mask MDYhms) in the do file when I use the my data set that is generate by briefcase with submission date like 3/2/2019 9:32.

The odkmeta do-file uses the Stata clock() function to convert dateTime strings to Stata %tc values. As part of that, the do-file specifies a "mask" that describes how the dateTime strings are formatted. See here for the Stata documentation about this mask.

The default mask that the do-file uses for dateTime strings is "MDYhms". However, you can use the `datetimemask' local macro at the top of the do-file to specify a different mask (see my post above). Your example value of 3/2/2019 9:32 seems like it might require the mask "MDYhm" (without the seconds component) or perhaps "DMYhm".

However, it seems unusual that your dateTime data does not include seconds. Because of that, I recommend the following steps:

1. Re-export the data from Briefcase.
2. Do not edit any CSV export (or even open it and re-save it) in Excel or any other software before running the odkmeta do-file: the do-file requires the raw data.
3. Next, try running the do-file without changing `datetimemask'.
4. If the do-file still results in an error, consider changing `datetimemask'. Consult the Stata documentation for the correct mask.

There is also a feature request to have Briefcase export date and time strings using ISO 8601, which I think would make things easier in cases like this by increasing standardization.

Hope this helps!