Changing string into integer in a column using lambda function doesn't work

Question

In my dataset there is a column that contains yes-no values, so I want to change them into 1-yes and 0-no using lambda function.

Code:

df['columnname'] = df['columnname'].apply(lambda a:1 if a == 'Yes' else 0)

The problem is all of data in my dataset turns into 0. Where is the error?

Is very hard to say without looking at the actual data, but using apply here seems like a misuse, you could simply use `df["columname"].eq("Yes")` — Dani Mesejo, Jul 18 '22 at 12:37
can you share the dataset please? upload on drive and share link here — Himanshu Poddar, Jul 18 '22 at 12:37
Don't use `apply`: `df['columnname'].eq('yes').astype(int)` or `df['columnname'].map({'yes': 1, 'no': 0})` — mozway, Jul 18 '22 at 12:37
right, the map function seems way more reasonable, thank you — gosiia, Jul 18 '22 at 12:40
@mozway why? both of them will eventually loop internally right? — Himanshu Poddar, Jul 18 '22 at 12:40
@HimanshuPoddar do the timing by yourself using `s = pd.Series(np.random.choice(['yes', 'no'], size=1_000_000))`. `.eq('yes').astype(int)` is ~4.5x faster, `map` is ~6.5x faster — mozway, Jul 18 '22 at 12:41

score 0 · Answer 1 · answered Jul 18 '22 at 12:38

0

You can use df.replace:

new_label = {'columnname': {'yes': 1, 'no': 0}}
df.replace(new_label, inplace=True)

answered Jul 18 '22 at 12:38

Jorgo

1 Answers1