Drop function in Pandas is changing original dataframe?

Question

I am using drop in pandas with an inplace=True set. I am performing this on a duplicate dataframe, but the original dataframe is also being modified.

df1 = df
for col in df1.columns:
   if df1[col].sum() > 1:
      df1.drop(col,inplace=True,axis=1)

This is modifying my 'df' dataframe and don't seem to understand why.

You don't know a very important concept. Pandas `DataFame` is a **mutable** type object. For better understanding check this out: [https://medium.com/@meghamohan/mutable-and-immutable-side-of-python-c2145cf72747] — oo00oo00oo00, Nov 27 '19 at 21:04
I do, I usually use the copy function as stated in another answer. Just didn't use it here. — user10939484, Nov 27 '19 at 21:06
"I am performing this on a duplicate dataframe" Your code says otherwise. You simply have two names for the same dataframe. — MisterMiyagi, Nov 27 '19 at 21:12
Does this answer your question? [Python Copy Through Assignment?](https://stackoverflow.com/questions/2438938/python-copy-through-assignment) — MisterMiyagi, Nov 27 '19 at 21:13

mcsoini · Accepted Answer · 2019-11-27T21:14:13.943

3

Use df1 = df.copy(). Otherwise they are the same object in memory.

However, it would be better to generate a new DataFrame directly, e.g.

df1 = df.loc[:, df.sum() <= 0]

edited Nov 27 '19 at 21:14

answered Nov 27 '19 at 21:02

mcsoini

Thanks, I don't know why I forgot that. – user10939484 Nov 27 '19 at 21:05
2

It'd be less memory to use `drop` without `inplace=True` and do the assignment in a single step – roganjosh Nov 27 '19 at 21:06

1 Answers1