mjhjmtu

I want to take the logarithm of every value in a pandas dataframe. I have tried this but it does not work:

#Reading data from excel and rounding values on 2 decimal places
import math
import pandas as pd

data = pd.read_excel("DataSet.xls").round(2)
log_data= math.log10(data)

It gives me this error:

TypeError: must be real number, not DataFrame

Do you have any idea what to do?

Use the numpy version, not math

import numpy as np

np.log10(df)

From what it seems math.log10 cannot handle neither pandas dataframes nor ndarrays.

So one option would be to go with numpy, which also includes a function to compute the base 10 logarithm, np.log10, and reconstruct the dataframe as pointed out in other solutions.

Or if you want to go with math.log10, and the same would apply to other functions that cannot be directly vectorized, you can use DataFrame.applymap to apply math.log10 to the dataframe elementwise. Do note however that this solution will be slower than a vectorized approach using np.log10.

Use case

Here's an example of how this could be done using DataFrame.applymap:

df = pd.DataFrame(np.random.randint(1,5,(6,6)), columns=list('abcdef'))

print(df)
 a b c d e f
0 3 4 1 1 2 1
1 4 4 4 3 4 1
2 4 3 3 1 4 1
3 3 4 1 3 1 1
4 1 2 3 4 2 1
5 1 3 3 1 4 3

df.applymap(math.log10)

 a b c d e f
0 0.477121 0.602060 0.000000 0.000000 0.30103 0.000000
1 0.602060 0.602060 0.602060 0.477121 0.60206 0.000000
2 0.602060 0.477121 0.477121 0.000000 0.60206 0.000000
3 0.477121 0.602060 0.000000 0.477121 0.00000 0.000000
4 0.000000 0.301030 0.477121 0.602060 0.30103 0.000000
5 0.000000 0.477121 0.477121 0.000000 0.60206 0.477121

For the numpy solution, you could take the np.log10 of the dataframe, and reconstruct it as:

pd.DataFrame(np.log10(data), index=df.index, columns=df.columns)

You may want to use the applymap method to apply math.log10 on the whole dataframe, here is the documentation.

You can test it:

df.applymap(math.log10)