I need to count occurrences of a certain value (let’s assume it’s 3) in a range of columns per each case. To do so I wrote a script as below:
import pandas as pd import numpy as np objsourcedf = pd.DataFrame({"a": [1, 2, 2], "b": [3, 1, 1], "c": [3, 2, 1], "d": [4, 3, 8]}) print(objsourcedf) objauxdf = objsourcedf.transpose() objauxdf.loc["counts"] = np.sum(objauxdf == 3) objsourcedf = objsourcedf.assign(counts=list(objauxdf.loc["counts"])) print(objsourcedf)
First print
is:
a b c d 0 1 3 3 4 1 2 1 2 3 2 2 1 1 8
Second:
a b c d counts 0 1 3 3 4 2 1 2 1 2 3 1 2 2 1 1 8 0
Even though it works fine I am pretty sure there is a more pythonic way to do so. By ‘pythonic’ I mean using native, concise pandas
feature and no looping through columns/rows. For example, in SPSS there is a simple count
command so regarding this objsourcedf
this line would be:
count counts = a b c d (3). execute.
Sadly, as a beginner in Python and pandas
I couldn’t find anything so I’m asking you if there’s a more simple way to get occurences?
Advertisement
Answer
I hope this qualifies at being “Pythonic”:
objsourcedf['count'] = objsourcedf.eq(3).sum(axis=1)