BUG: DataFrame[interval[td64]] transpose fails

[jbrockmendel]

tdi = pd.timedelta_range("0 Days", "3 Days")
ii = pd.IntervalIndex.from_breaks(tdi)
ii = ii.insert(-1, np.nan)
df = pd.DataFrame(ii)

>>> df.T
[...]
TypeError: Cannot convert Float64Index to dtype timedelta64[ns]; integer values are required for conversion

This would be solved by having 2D EAs.

[sukriti1]

take

[weikhor]

@jbrockmendel Hi I see this from above This would be solved by having 2D EAs. . May I know what is meaning of 2D EAs?

[jbrockmendel]

2D EAs means making all ExtensionArrays support 2D. There is not consensus on doing this, which seems unlikely to change.

[weikhor]

2D EAs means making all ExtensionArrays support 2D. There is not consensus on doing this, which seems unlikely to change.

Ok thank