CODEWITHSUNDEEP
pythonpandasnumpydataframe
Mathew Carroll
Mathew Carroll

Reputation: 377

Python (pandas): How to multiply all floats in DataFrame by constant

if I have a DataFrame consisting of some columns that only consist of strings, and other columns that only consist of numeric types (i.e. float or int), how to I multiply each float or int element in the DataFrame by some constant?

For example, if the constant is 3, and I have this data frame:

import numpy as np
import pandas as pd

np.random.seed(seed=9876)
df1 = pd.DataFrame(['a']*3+['b'])
df2 = pd.DataFrame(['x','y','z', 't'])
df3 = pd.DataFrame(np.round(np.random.randn(4,2),2)*10)
df = pd.concat([df1, df2, df3], axis = 1)
df.columns = ['ind', 'x1', 'x2','x3']
df = df.set_index('ind')
print(df)
    x1    x2    x3
ind               
a    x   3.9 -10.9
a    y   2.1   3.2
a    z  -9.3   0.3
b    t -11.1  -1.2

The output should look like:

    x1    x2    x3
ind               
a    x   11.7   -32.7
a    y   6.3    9.6
a    z  -27.9   0.9
b    t  -33.3   -3.6

One idea is to use pd.concat([df['x1'], df.loc[:,['x2', 'x3']].mul(3)], axis = 1), but I won't necessarily know ex-ante which are the string vs numeric columns when applying this to numerous data frames. It also seems inefficient. I'm sure a better method exists but I'm yet to come up with one.

Upvotes: 0

Views: 1229

Answers (1)

jezrael
jezrael

Reputation: 862751

Select only columns with numeric by DataFrame.select_dtypes and then multiple only them:

cols = df.select_dtypes(np.number).columns
print (cols)
Index(['x2', 'x3'], dtype='object')

df[cols] = df[cols].mul(3)

print (df)
    x1    x2    x3
ind               
a    x  11.7 -32.7
a    y   6.3   9.6
a    z -27.9   0.9
b    t -33.3  -3.6

Upvotes: 2

Related Questions