Get rid of NaT and duplicates from a pandas dataFrame to obtain a series of datetime values

  dataframe, pandas, python, python-datetime

I have a dataframe that looks like shown picture enter image description here

Dataframe shape is (1944,900).
Each row of the dataframe has one value (it might be repeated multiple times based on the row index). I need to extract a list of 1944 numbers each representing valid value from each row (excluding NaT and duplicate values).

Any ideas on this?

Source: Python Questions