Hi,
I’m using apply to do value_counts() on two columns at once. I, however, want to set dropna=False.
This is not possible as I cannot supply any additional arguments for value_counts once I use it in apply.
columns = ['dissatisfaction','job_dissatisfaction']
print(tafe_resignations[columns].apply(pd.value_counts(dropna=False))
Is there a work-around for this, or should I do value_counts for both columns seperately?
Hi Bruno,
I tried
print(tafe_resignations[columns].apply(pd.DataFrame.value_counts))
But then I get an error saying
Am I not looking in the right direction with your hint? 
Review the examples section in the link I sent. If that’s not enough reach out again. Also please include a link to the screen in your original post.
Hi Bruno, I checked the example section but cant come to a solution.
I currently have this:
The following does what I want:
print(tafe_resignations['dissatisfaction'].value_counts(dropna=False))
print('\n')
print(tafe_resignations['job_dissatisfaction'].value_counts(dropna=False))
But i’d like to do it in one step. Therefore I tried:
print(tafe_resignations[columns].apply(pd.value_countss))
Which does exactly what I want but won’t allow argument dropna = false.
When I tried:
print(tafe_resignations.value_counts(dropna=False)
I get an error saying that value_counts() got an unexpected argument ‘dropna’ even though this argument is mentioned in the documentation.
Nevertheless, even if it would allow the dropna=False I do not think it does the trick as it returns a series of counts of unique rows.
You could try a lambda function:
.apply(lambda x: x.value_counts(dropna=False))
Works like a charm! Thank you @vik.
I’ve seen lambda functions a couple of times but thought to leave it for now as I assumed it will be part of the Data Analyst with Python path I’m taking.
Does anybody know why I get the unexpected argument (dropna=False) error with DataFrame.value_counts()?
Because calling .apply(pd.value_counts(dropna=False)) immediately ran the value_counts function, versus passing the reference into the apply method. When you instead did .apply(pd.value_counts), then it passed in the function reference, and the apply method ran the function across each column. It’s the difference between a = sum and a=sum([1,2]).
