I have four datasets with overlapping datetime columns. I want to produce an analysis showing just the date columns as a single dataframe to reveal the overlapping dates. Similar to this,
| A | B | C |
| 2017-05-01 | NAN | 2017-05-01 |
| 2017-05-02 | NAN | 2017-05-02 |
| 2017-05-03 | 2017-05-03 | 2017-05-03 |
| ... | ... | ... |
| 2017-05-07 | 2017-05-07 | NAN |
I mocked this up by randomly selecting from pd.date_range to produce three dataframes with size (7,1). The datasets could be of different lengths or some might have gaps in their series.
I tried join,
dfZ.set_index['Date']
X = dfA.join(dfB, how='outer', on='Date1', rsuffix='_B').join(dfC, how='outer', on='Date3', rsuffix='_C')
but this produces an error: #TypeError: 'method' object is not subscriptable
Concat will join the tables on the correct axis,
X = pd.concat([dfA, dfB, dfC])
But they're not aligned, so the size of the df is (21,3) and not (7,3).