Hi @insongkim, thanks so much for this amazing package and contribution!
I was playing around with the package for a project of mine, in which the time variable is a factor derived from a date. So all the factor labels are dates, not integers.
It turns out this causes a bug that I've tracked down. Should be an easy fix I think.
See here for a minimum working example using the third example (observational 2-way DiD) from the ?wfe documentation:
Hi @insongkim, thanks so much for this amazing package and contribution!
I was playing around with the package for a project of mine, in which the time variable is a factor derived from a date. So all the factor labels are dates, not integers.
It turns out this causes a bug that I've tracked down. Should be an easy fix I think.
See here for a minimum working example using the third example (observational 2-way DiD) from the
?wfe
documentation:The error occurs here, lines 466-473 of the script with
wfe()
in the version currently on CRAN:The last line is valued zero because
idxall
assumes integers for the date variable, and therefore finds no matches withidx
.Hope this is helpful for future reference!