Convert TimeIndex of dataframe to DateTimeIndex in place

Question

Coming from a MultiIndex:

I can select the TimeIndex using get_level_values

hist_vol_to_display.index.get_level_values('ts_timeonly')

So all of the rows have happend on a specific date (e.g. 2019-09-16).

What is an efficient way to replace the TimeIndex with a DateTimeIndex?

Can I replace the Index inplace or do I need to work around this using a new column? Thanks!

Answer 1

You can use MultiIndex.set_levels:

x  = pd.to_datetime('2019-09-16 ' + hist_vol_to_display.index.levels[1].astype(str))
hist_vol_to_display.index = hist_vol_to_display.index.set_levels(x, level=1)