df[“column”] and df.column The difference between

import numpy as np
import pandas as pd

columns = ["name","id"]
names = ["wangfang","zhangsan","jeff","peter","hell"]
ids = [i** 2 for i in range(5)]
df = pd.DataFrame(data=np.array([names,ids]).T,columns=columns)

df

nameid0wangfang01zhangsan12jeff43peter94hell16

• adopt df.column and df[“column”] The column values can be obtained in any of the following ways
• however df.column There are restrictions on the way ： That's it column The name of must be valid

Valid column name presentation

df.name

0 wangfang
1 zhangsan
2 jeff
3 peter
4 hell
Name: name, dtype: object

df["name"]

0 wangfang
1 zhangsan
2 jeff
3 peter
4 hell
Name: name, dtype: object

Invalid column name Demo

series_ = pd.Series([24,38,66,52,31])
df[" Age ~ yes "]= series_

df

nameid Age ~ yes 0wangfang0241zhangsan1382jeff4663peter9524hell1631

# You can't pass at this time df.column To get the data of this column 
df[" Age ~ yes "]

0 24
1 38
2 66
3 52
4 31
Name: Age ~ yes , dtype: int64

df.column Columns cannot be created or modified df

# In this way, you can create , You can also batch modify column values 
df["add"] = [1,2,3,4,5]
df

nameid Age ~ yes add0wangfang02411zhangsan13822jeff46633peter95244hell16315

df["add"][0]= 100
df

D:\Anaconda3\envs\data\lib\site-packages\ipykernel_launcher.py:1: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
"""Entry point for launching an IPython kernel.

nameid Age ~ yes add0wangfang0241001zhangsan13822jeff46633peter95244hell16315

# What you get with attributes is a copy , Can not be used against the original df Make changes , You cannot create new columns 
# Cannot be modified directly df
df.add = [1,6,4,2,2]
df.add[0] = 10
# Cannot create new column 
df.add_ = [0,0,0,0,0]
df

nameid Age ~ yes add0wangfang0241001zhangsan13822jeff46633peter95244hell16315

index Repeatable but not modifiable

index It's repeatable

data = np.array([1,2,4,5,8])
series = pd.Series(data=data,index=['a','a','b','c','c'])

series

a 1
a 2
b 4
c 5
c 8
dtype: int32

series["a"]

a 1
a 2
dtype: int32

index It cannot be modified separately

series.index[0] = "4"

---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-22-a68b0d9a21b2> in <module>
----> 1 series.index[0] = "4"
D:\Anaconda3\envs\data\lib\site-packages\pandas\core\indexes\base.py in __setitem__(self, key, value)
4082
4083 def __setitem__(self, key, value):
-> 4084 raise TypeError("Index does not support mutable operations")
4085
4086 def __getitem__(self, key):
TypeError: Index does not support mutable operations

index It can be replaced in batches

• But be sure to give index The length of the array is consistent with the original data
• adopt reindex Method does not need to keep the data length consistent ; Because items that do not exist will follow NAN Value filling

series.index = [str(i) for i in range(5)]

series

0 1
1 2
2 4
3 5
4 8
dtype: int32

series.reindex([str(i) for i in range(6)])

0 1.0
1 2.0
2 4.0
3 5.0
4 8.0
5 NaN
dtype: float64

• Of course NaN Value can be used method The method in , choice ffill Is to fill in with the previous value

series.reindex([str(i) for i in range(6)],method="ffill")

0 1
1 2
2 4
3 5
4 8
5 8
dtype: int32

reindex Common parameters of

• method: The way of interpolation “ffill” “bfill” Fill in the preceding item and fill in the following item
• fill_value: Fill with a special value
• limit: The maximum dimension gap to be filled when filling the front or rear items
• tolerance: When the preceding or subsequent items are filled , The maximum dimension gap under the inaccurate match to be filled
• level
• copy