Closed frreiss closed 3 years ago
The issue here is that the Series uses the Pandas Datetime64Formatter
which converts the values to a DatetiemIndex
that expects to be 1-dimensional. I will look some more for a possible workaround or upstream fix.
Submitted a fix upstream to Pandas at https://github.com/pandas-dev/pandas/pull/38391
Code to reproduce:
Expected result: Display a 3x5 matrix of timestamps
Actual result: Crash from inside a Pandas routine that should only be called for 1-D arrays of timestamps. Stack trace follows.