pandas append same series to each column

Question

Consider the dataframe df

df = pd.DataFrame(np.random.rand(5, 3),
                  ['p0', 'p1', 'p2', 'p3', 'p4'],
                  ['A', 'B', 'C'])
df

And the mean of every row:

dm = df.mean(1)
dm.index = ['m0', 'm1', 'm2', 'm3', 'm4']
dm

m0    0.352396
m1    0.606469
m2    0.643022
m3    0.560809
m4    0.776058
dtype: float64

How do I append this series of means to every column of df. I expect the results to look like:

Also, since this will be applied at scale, time is of the essence.

What I used to generate the expected output is:

pd.concat([df, pd.DataFrame({c: dm for c, i in df.iteritems()})])

---

Timing (small scale)

Timing (large scale)

Answer 1

You can use double concat:

print (pd.concat([dm] * df.shape[1], axis=1, keys=df.columns))
           A         B         C
m0  0.823788  0.823788  0.823788
m1  0.615354  0.615354  0.615354
m2  0.606740  0.606740  0.606740
m3  0.386629  0.386629  0.386629
m4  0.637147  0.637147  0.637147

print (pd.concat([df, pd.concat([dm] * df.shape[1], axis=1, keys=df.columns)]))
           A         B         C
p0  0.789966  0.699837  0.981560
p1  0.415609  0.469310  0.961144
p2  0.920938  0.476615  0.422665
p3  0.323782  0.805231  0.030874
p4  0.761674  0.361134  0.788632
m0  0.823788  0.823788  0.823788
m1  0.615354  0.615354  0.615354
m2  0.606740  0.606740  0.606740
m3  0.386629  0.386629  0.386629
m4  0.637147  0.637147  0.637147

For creating appended df is possible use numpy repeat and numpy.newaxis:

x = dm.values

print (pd.DataFrame(np.repeat(x[:, np.newaxis], df.shape[1], 1), 
                    columns=df.columns,
                    index=dm.index))
           A         B         C
m0  0.399837  0.399837  0.399837
m1  0.890191  0.890191  0.890191
m2  0.580747  0.580747  0.580747
m3  0.354032  0.354032  0.354032
m4  0.329108  0.329108  0.329108

print(pd.concat([df, pd.DataFrame(np.repeat(x[:, np.newaxis], df.shape[1], 1), 
                    columns=df.columns,
                    index=dm.index)]))

           A         B         C
p0  0.087337  0.375891  0.736282
p1  0.777897  0.932047  0.960629
p2  0.945546  0.062647  0.734047
p3  0.247740  0.582076  0.232282
p4  0.078683  0.869736  0.038905
m0  0.399837  0.399837  0.399837
m1  0.890191  0.890191  0.890191
m2  0.580747  0.580747  0.580747
m3  0.354032  0.354032  0.354032
m4  0.329108  0.329108  0.329108    

EDIT1:

Another solution for creating new df with numpy.tile:

dm2 = pd.DataFrame(np.tile(dm.values[:, None], (1, df.shape[1])), dm.index, df.columns)
df.append(dm2)