Joining two Pandas Dataframe

Question

Please, given the two dataframes:

DF1:

A B a1 b1 a2 b2 a3 b3

DF2:

C1 C2 C3 0 0 1

I would like to do the following DF1+DF2 producing the following:

 A B C1 C2 C3 a1 b1 0 0 1 a2 b2 0 0 1 a3 b3 0 0 1

it is not clear to me how I should use one of Merge, Join or concatenate. Please, any help would be very appreciated.c.

Possible duplicate of Pandas: Join dataframes on selected columns — raam86
– raam86, Commented Nov 6, 2017 at 18:50

Scott Boston · Accepted Answer · 2017-11-06 18:55:39Z

Let's use some unpacking and forward fill nas:

DF1.assign(**DF2).ffill()

OR, let's create a dummy key to do a cartesian join, and drop the dummy key.

DF1.assign(key=1).merge(DF2.assign(key=1), on='key').drop('key',axis=1)

Output:

 A B C1 C2 C3 0 a1 b1 0.0 0.0 1.0 1 a2 b2 0.0 0.0 1.0 2 a3 b3 0.0 0.0 1.0

BENY · Accepted Answer · 2017-11-06 18:52:13Z

4

pd.concat + ffill

pd.concat([df1,df2],1).ffill() Out[1188]: A B C1 C2 C3 0 a1 b1 0.0 0.0 1.0 1 a2 b2 0.0 0.0 1.0 2 a3 b3 0.0 0.0 1.0

answered Nov 6, 2017 at 18:52

324k22 gold badges176 silver badges250 bronze badges

3 Comments

does not solved the problem. It add NaN values. The solution is provided by Scott Boston above.

@CarloAllocca .....what is your problem ? So you do not want any NaN during the calculation ?

Thanks for replying Wen. Yes, the problem was there were some additional NaN values at the first row which I do not want.

MaxU - stand with Ukraine · Accepted Answer · 2017-11-06 19:22:07Z

1

If df2 contains only one row we can do it this way:

In [30]: df1.assign(**df2.iloc[0].to_dict()) Out[30]: A B C1 C2 C3 0 a1 b1 0 0 1 1 a2 b2 0 0 1 2 a3 b3 0 0 1

answered Nov 6, 2017 at 19:22

212k37 gold badges402 silver badges436 bronze badges

Yeah... I do like that way better it safer with one row.