Cfrgtkky

In [66]: data

Out[66]: 

   col1 col2 label

0   1.0    a     c

1   2.0    b     d

2   3.0    c     e

3   0.0    d     f

4   4.0    e     0

5   5.0    f     0



In [67]: data.label

Out[67]: 

0      c

1      d

2    NaN

3      f

4    NaN

5    NaN

Name: col2, dtype: object



In [68]: data['label']

Out[68]: 

0    c

1    d

2    e

3    f

4    0

5    0

Name: label, dtype: object

Why data.label and data['label'] showing different results?

The difference between these two is related to assignment. with data.label you cannot assign the values to column.

data.label is to access the attributes and data["label"] is to assign the values.

Also if you have spaces in your column name, for example df['label name'], while using data.label name will through an error.

For more information see this Answer link

The big difference I've noticed is assignment.

import random

import pandas as pd



s = "SummerCrime|WinterCrime".split("|")

j = {x: [random.choice(["ASB", "Violence", "Theft", "Public Order", "Drugs"]) for j in range(300)] for x in s}

df = pd.DataFrame(j)

df.FallCrime = [random.choice(["ASB", "Violence", "Theft", "Public Order", "Drugs"]) for j in range(300)]

Gives: UserWarning: Pandas doesn't allow columns to be created via a new attribute name

However, there are also docs associated with this: https://pandas.pydata.org/pandas-docs/stable/indexing.html#attribute-access

Which has the following warnings which may be related to your problem:

- You can use this access only if the index element is a valid Python

   identifier, e.g. s.1 is not allowed. 

 - The attribute will not be available if it

   conflicts with an existing method name, e.g. s.min is not allowed.

 - Similarly, the attribute will not be available if it conflicts with

   any of the following list: index, major_axis, minor_axis, items. In

   any of these cases, standard indexing will still work, e.g. s['1'],

   s['min'], and s['index'] will access the corresponding element or

   column.

They go on to say:

You can use attribute access to modify an existing element of a Series or column of a

DataFrame, but be careful; if you try to use attribute access to create a new column,

it creates a new attribute rather than a new column.

**In 0.21.0 and later, this will raise a UserWarning** 

(So it's possible you did this without realizing)