熊猫：在列中查找最大字符串值的索引

Question

I have some time values that are in string format. I would like to return the value of the most recent or "max" time in a column. I noticed that argmax and idmax don't work for strings. I would like to avoid converting them to find the index of the max value. Is there a way to return the index of the max string value?

df['Start_time'].max()

returns: '2017-05-02 15:47:21'

df.loc[df['Start_time'].idxmax()]

returns: ValueError: could not convert string to float: '2017-01-26 16:33:16'

Answer 1

If dont want converting to datetime , you can use boolean indexing with comparing column with max value:

print (df[df['Start_time'] == df['Start_time'].max()])

And for index is possible use:

df.index[df['Start_time'] == df['Start_time'].max()][0]

Or:

df[df['Start_time'] == df['Start_time'].max()].index[0] 

Answer 2

Source DF:

In [60]: df
Out[60]:
            Start_time
0  2017-05-01 16:16:16
1  2017-05-02 15:47:21
2  2017-05-01 10:10:10

Solution 1:

In [61]: df.iloc[df['Start_time'].values.argmax()]
Out[61]:
Start_time    2017-05-02 15:47:21
Name: 1, dtype: object

Solution 2:

In [62]: df.loc[pd.to_datetime(df['Start_time']).idxmax()]
Out[62]:
Start_time    2017-05-02 15:47:21