df['colimn_name'] vs df.loc[:, 'colimn_name']

Question

I would like more info. on the answer to the following question:

df[‘Name’] and 2. df.loc[:, ‘Name’], where:

df = pd.DataFrame(['aa', 'bb', 'xx', 'uu'], [21, 16, 50, 33], columns = ['Name', 'Age'])

Choose the correct option:

1 is the view of original dataframe and 2 is a copy of original
dataframe
2 is the view of original dataframe and 1 is a copy of original dataframe
Both are copies of original dataframe
Both are views of original dataframe

I found more than one answer online but not sure. I think the answer is number 2 but when i tried x = df['name'] then x[0] = 'cc' then print(df) I saw that the change appeared in the original dataframe. So how the changed appeared in the original dataframe although I also got this warining: A value is trying to be set on a copy of a slice from a DataFrame

I just want to know more about the difference between the two and weather one is really a copy of the original dataframe or not. Thank you.

Linked with [What is the difference between using loc and using just square brackets to filter for columns in Pandas/Python?](https://stackoverflow.com/a/48411543/11154841). — questionto42, Feb 28 '23 at 16:51

BENY · Accepted Answer · 2020-07-18T16:27:37.660

2

Both of them are VIEW only, if you need a copy we need

df['Name'].copy(). 
df.loc[:, 'Name'].copy()

loc here can give you access to index and column at the same time, also avoid chain slice

df.loc[cond1, cond2]

Notice above loc can also pass condition rather than only column names

df.loc[:, df.columns=='Name']
#df[df.columns=='Name'] this will return error

Example for Div

df=pd.DataFrame([[1,2,3],[2,3,4]])
df
   0  1  2
0  1  2  3
1  2  3  4
df[df.columns==0]

Traceback (most recent call last):

edited Jul 18 '20 at 16:27

answered Jul 18 '20 at 16:07

BENY

317,841
20
164
234

Condition can also be given in `df[]` – Divyessh Jul 18 '20 at 16:11
@DivyesshMaheshwari you sure ? – BENY Jul 18 '20 at 16:13
@DivyesshMaheshwari if condition that is index slice in [] – BENY Jul 18 '20 at 16:14
I am sure and any condition can be placed brother – Divyessh Jul 18 '20 at 16:15
1

@DivyesshMaheshwari try the example – BENY Jul 18 '20 at 16:27
The condition can be given as `df[df['0'] > 1]` – Divyessh Jul 18 '20 at 17:01
@DivyesshMaheshwari this is index slice condition ...I mean the column axis=1 – BENY Jul 18 '20 at 17:02
Yes that is what loc is for – Divyessh Jul 18 '20 at 17:05
The df[] has its own condition – Divyessh Jul 18 '20 at 17:05
Let us [continue this discussion in chat](https://chat.stackoverflow.com/rooms/218103/discussion-between-divyessh-maheshwari-and-yoben-s). – Divyessh Jul 18 '20 at 17:05

score 0 · Answer 2 · answered Jul 18 '20 at 16:04

0

Both are the views of original dataframe One can be used to add more columns in dataframe and one is used for specifically getting a view of a cell or row or column in dataframe.

answered Jul 18 '20 at 16:04

Divyessh

2,540
1
7
24

df['colimn_name'] vs df.loc[:, 'colimn_name']

2 Answers2