What's the fastest way to compare datetime in pandas?

Question

I have two big csv files with different number of rows which I am importing as follows:

tdata = pd.read_csv(tfilepath, sep=',', parse_dates=['date_1'])
print(tdata.iloc[:, [0,3]])

       TBA    date_1
0       0   2010-01-04
1       9   2010-01-05
2       0   2010-01-06    
3       8   2010-01-07
4       0   2010-01-08
5       0   2010-01-09

pdata = pd.read_csv(pfilepath, sep=',', parse_dates=['date_2'])
print(pdata.iloc[:, [0,3]])

       TBA    date_2
0       3   2011-01-04
1       5   2010-01-09
2       0   2012-02-03    
3       9   2010-03-17
4       1   2010-11-08
5       2   2010-01-05

Now I want to replace TBA in the first dataframe with corresponding TBA in the second dataframe where the dates match. The default value would be 0. So I am iterating through rows as follows:

 for i, row1 in tdata.iterrows():
     for j, row2 in pdata.iterrows():
         if row1['date_1'] == row2['date_2']:
             tdata.loc[i, 'TBA'] = row2['TBA']
             break
         else:
             tdata.loc[i, 'TBA'] = 0

Problem is this takes very long (around 11 minutes). I want to compare one csv with 160 other csv and further run some tree based models. I am a newbee with little coding background! Pardon me if this is a 'dirty' way. Any help would be appreciated. Thanks!

Answer 1

If you call set_index on pdata to date_2 then you can pass this as the param to map and call this on tdata['date_1'] column and then fillna :

In [51]:    
tdata['TBA'] = tdata['date_1'].map(pdata.set_index('date_2')['TBA'])
tdata.fillna(0, inplace=True)
tdata

Out[51]:
   TBA      date_1
0    0  2010-01-04
1    2  2010-01-05
2    0  2010-01-06
3    0  2010-01-07
4    0  2010-01-08
5    5  2010-01-09