Python DataFrame取行

关注根据某列series的值区间，取行问题。
根据行的index区间位置，必须先知道index区间，比较基础
代码准备：
环境平台：Python 3.7.1 -IDLE Shell

>>>import pandas as pd
>>> df = pd.DataFrame({'Name': ['Tom', 'Jim', 'Lily'], 'Age': [20, 18, 22], 'Gender': ['Male', 'Male', 'Female']})

注：该例子数据形式来自：https://www.python100.com/html/116332.html

index区间取行

示例1：提取索引名=‘1’的那一行，返回一个行Series

>>> row = df.loc[1]#按索引名提取，当使用自动生成的索引时，索引名与索引号相同
>>> row
Name       Jim
Age         18
Gender    Male
Name: 1, dtype: object
>>> type(row)
<class 'pandas.core.series.Series'>

注：

df.iloc[:] #按索引(号)提取

示例2：

row = df.loc[0:1]
>>> row
  Name  Age Gender
0  Tom   20   Male
1  Jim   18   Male
>>> row = df.iloc[0:1]
>>> row
  Name  Age Gender
0  Tom   20   Male
>>> row = df.loc[0:0]
>>> row
  Name  Age Gender
0  Tom   20   Male
>>> type(row)
<class 'pandas.core.frame.DataFrame'>

注意：两种提取的区间有区别：按索引（号）提取的区间为：[0,1)

>>> row = df.iloc[0:0]
>>> row
Empty DataFrame
Columns: [Name, Age, Gender]
Index: []

列值区间条件取行

>>> row = df.loc[df['Age'] > 20].iloc[0]['Name']
>>> row
'Lily'
>>> 

上语句的含义是：需要从dataframe：df.loc[df[‘Age’] > 20]中提取索引为0的行Series的‘name’值

（1）列值区间基本表达方式

示例1：

>>> row = df.loc[df['Age'] > 18]
>>> row
   Name  Age  Gender
0   Tom   20    Male
2  Lily   22  Female
>>> 

注：超过区间，不会产生错误，返回：

>>> row = df.loc[df['Age'] > 23]
>>> row
Empty DataFrame
Columns: [Name, Age, Gender]
Index: []

（2）多条件组合表达方式

示例2：

>>> row = df.loc[(df['Age'] >= 18)&(df['Name'] == 'Lily')]
>>> row
   Name  Age  Gender
2  Lily   22  Female
>>> 

如果条件为False则返回的dataframe为Empty：

>>> row = df.loc[(df['Age'] >= 18)&(df['Name'] == 'tongzhi')]#'tongzhi'不存在原dataframe
>>> row
Empty DataFrame
Columns: [Name, Age, Gender]
Index: []
>>> 

当然也可以：用’|'关系操作符：

>>> row = df.loc[(df['Age'] >= 18)|(df['Name'] == 'Jim')]
>>> row
   Name  Age  Gender
0   Tom   20    Male
1   Jim   18    Male
2  Lily   22  Female
>>> 

注：还可以关系:~ 非

（3）函数条件表达方式

可以使用lambda或自定义函数（返回bool）选择符合返回条件的行，如：

>>> x='Jim'
>>> row = df.loc[lambda x:x['Name'] == 'Jim']
>>> row
  Name  Age Gender
1  Jim   18   Male
>>> 

datafame接受的几个过滤函数

（1）isin函数：

df[df[“column_name”].isin(li)] (# li = [20, 25, 27] 或 li = np.arange(20, 30))
根据从isin函数传入的列表(li)，筛选出与列表中包含的数值或字符串相同的数据记录, 用法有点类似sql中的"in"

(2) query函数：

df.query(“(column_name1 == ‘str1’) & (column_name2 == ‘str2’)”)
根据query中引入的不同字段（str1，str2等）和条件，筛选出同时能满足这些要求的数据记录

(3) contains函数：

df[df[“column_name”].str.contains(“str”)]
筛选出所有含有(str)的数据记录, 用法类似于sql中的"contains"

以上参考了：链接：https://blog.csdn.net/weixin_45914452/article/details/120585861

错误条件格式：

示例1：

>>> row = df.loc[(18<=df['Age'] <= 22)]
Traceback (most recent call last):
  File "<pyshell#56>", line 1, in <module>
    row = df.loc[(18<=df['Age'] <= 22)]
  File "D:\Program Files\Python371\lib\site-packages\pandas\core\generic.py", line 1538, in __nonzero__
    f"The truth value of a {type(self).__name__} is ambiguous. "
ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().
>>> 

示例2：

>>> row = df.loc[df['Age'] >= 18 & df['Age'] <= 22]
Traceback (most recent call last):
  File "<pyshell#38>", line 1, in <module>
    row = df.loc[df['Age'] >= 18 & df['Age'] <= 22]
  File "D:\Program Files\Python371\lib\site-packages\pandas\core\generic.py", line 1538, in __nonzero__
    f"The truth value of a {type(self).__name__} is ambiguous. "
ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().