Get columns names from one data frame and add them as empty columns in another data frame in pandas

Question

I have one data frame (df1) with 5 columns and another (df2) with 10 columns. I want to add columns from df2 to df1, but only columns names (without values). Also, I want to do the same with adding columns without values from df1 to df2.

Here are the data frames:

df1

  A      B     C      D     E     
  1      234   52     1     54
  54     23    87     5     125
  678    67    63     8     18
  45     21    36     5     65
  8      5     24     3     13

df2

  F      G     H      I     J      K      L     M      N     O     
  12     34    2      17    4      19     54    7      58    123
  154    3     7      53    25     2      47    27     84    6
  78     7     3      82    8      56     21    29     547   1

And I want to get this:

df1

  A      B     C      D     E      F      G      H      I      J      K      L      M      N      O     
  1      234   52     1     54
  54     23    87     5     125
  678    67    63     8     18
  45     21    36     5     65
  8      5     24     3     13

And I want to get this:

df2

  A       B       C       D      E      F      G     H      I     J      K      L     M      N     O     
                                        12     34    2      17    4      19     54    7      58    123
                                        154    3     7      53    25     2      47    27     84    6
                                        78     7     3      82    8      56     21    29     547   1

I tried with df.columns.values and got the array of columns names, but then I have to apply them as data frame columns and give them empty values, and the way that I am doing now has too many lines of code, and I just wonder is it some easier way to do that? I will appreciate any help.

jezrael · Accepted Answer · 2018-12-04 07:23:54Z

6

Use Index.union with DataFrame.reindex:

cols = df1.columns.union(df2.columns)
#if order is important
#cols = df1.columns.append(df2.columns)

df1 = df1.reindex(columns=cols)
df2 = df2.reindex(columns=cols)

print (df1)
     A    B   C  D    E   F   G   H   I   J   K   L   M   N   O
0    1  234  52  1   54 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
1   54   23  87  5  125 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
2  678   67  63  8   18 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
3   45   21  36  5   65 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
4    8    5  24  3   13 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN

print (df2)
    A   B   C   D   E    F   G  H   I   J   K   L   M    N    O
0 NaN NaN NaN NaN NaN   12  34  2  17   4  19  54   7   58  123
1 NaN NaN NaN NaN NaN  154   3  7  53  25   2  47  27   84    6
2 NaN NaN NaN NaN NaN   78   7  3  82   8  56  21  29  547    1

If same index values in both DataFrames is possible use DataFrame.align:

print (df1)
     A    B   C  D    E
0    1  234  52  1   54
1   54   23  87  5  125
2  678   67  63  8   18

df1, df2 = df1.align(df2)
print (df1)
     A    B   C  D    E   F   G   H   I   J   K   L   M   N   O
0    1  234  52  1   54 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
1   54   23  87  5  125 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
2  678   67  63  8   18 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN

print (df2)
    A   B   C   D   E    F   G  H   I   J   K   L   M    N    O
0 NaN NaN NaN NaN NaN   12  34  2  17   4  19  54   7   58  123
1 NaN NaN NaN NaN NaN  154   3  7  53  25   2  47  27   84    6
2 NaN NaN NaN NaN NaN   78   7  3  82   8  56  21  29  547    1

edited Dec 4, 2018 at 7:23

answered Dec 4, 2018 at 7:04

jezrael

868k103 gold badges1.4k silver badges1.3k bronze badges

Sign up to request clarification or add additional context in comments.

5 Comments

U13-Forward Over a year ago

Nice. +1, also can use .fillna('') to replace with empty as OP's desired output.

slobokv83 Over a year ago

@jezrael Is it possible to keep the same order of columns, not sorted? For example, if in df2 the order of columns is like this: J, F, O, G, M, H, N, L, K, I. What if I want to keep that order of columns? Not to use reindex?

jezrael Over a year ago

@U9-Forward - Thank you. Yes, if OP need it.

U13-Forward Over a year ago

@jezrael Happy to add :-)

jezrael Over a year ago

@slobokv83 - if order is important then use cols = df1.columns.append(df2.columns), added to answer.

Collectives™ on Stack Overflow

Get columns names from one data frame and add them as empty columns in another data frame in pandas

1 Answer 1

5 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

5 Comments

Your Answer

Sign up or log in

Post as a guest

Related