Pandas数据处理——盘点那些常用的函数(下)
继上一篇文章Pandas数据处理——盘点那些常用的函数(上),这篇文章整理了剩下的一些Pandas常见方法,整体难度会比上一篇文章中的大一点,但还是比较容易理解的。话不多说,直接进入正题。
用于演示的数据如下:
In [11]: data
Out[11]:
company gender salary age
0 B female 30 40.0
1 A female 36 31.0
2 B female 35 28.0
3 B female 9 18.0
4 B female 16 43.0
5 A male 46 22.0
6 B female 15 28.0
7 B female 33 40.0
8 C male 19 32.0
.astype()
作用对象:Series
和DataFrame
主要用途:修改字段的数据类型,数据量大的情况下可用于减小数据占用的内存,多用于Series
。
用法:
# 把age字段转为int类型
In [12]: data["age"] = data["age"].astype(int)
In [13]: data
Out[13]:
company gender salary age
0 B female 30 40
1 A female 36 31
2 B female 35 28
3 B female 9 18
4 B female 16 43
5 A male 46 22
6 B female 15 28
7 B female 33 40
8 C male 19 32
.rename()
作用对象:Series
,DataFrame
(大多数情况下)
主要用途:多用于修改DataFrame
的列名
主要参数:
- columns (dict-like or function)
指定要修改的列名以及新的列名,一般以字典形式传入 - inplace (boolean, default False)
是否作用于原对象
用法:
# 将'age'更改为员工编号'number',并作用于原对象
In [15]: data.rename(columns={'age':'number'},inplace=True)
In [16]: data
Out[16]:
company gender salary number
0 B female 30 40
1 A female 36 31
2 B female 35 28
3 B female 9 18
4 B female 16 43
5 A male 46 22
6 B female 15 28
7 B female 33 40
8 C male 19 32
.set_index()
作用对象:DataFrame
主要用途:将DataFrame
中的某一(多)个字段设置为索引
用法:
In [19]: data.set_index('number',inplace=True)
In [20]: data
Out[20]:
company gender salary
number
40 B female 30
31 A female 36
28 B female 35
18 B female 9
43 B female 16
22 A male 46
28 B female 15
40 B female 33
32 C male 19
.reset_index()
作用对象:Series
,DataFrame
主要用途:重置索引,默认重置后的索引为0~len(df)-1
主要参数:
- drop (boolean, default False)
是否丢弃原索引,具体看下方演示 - inplace (boolean, default False)
是否作用于原对象
用法:
# drop = True,重置索引,并把原有的索引丢弃
In [22]: data.reset_index(drop=True)
Out[22]:
company gender salary
0 B female 30
1 A female 36
2 B female 35
3 B female 9
4 B female 16
5 A male 46
6 B female 15
7 B female 33
8 C male 19
# drop = False,重置索引
# 原索引列'number'作为新字段进入DataFrame
In [23]: data.reset_index(drop=False,inplace=True)
In [24]: data
Out[24]:
number company gender salary
0 40 B female 30
1 31 A female 36
2 28 B female 35
3 18 B female 9
4 43 B female 16
5 22 A male 46
6 28 B female 15
7 40 B female 33
8 32 C male 19
.drop_duplicates()
作用对象:Series
,DataFrame
主要用途:去掉重复值,作用和SQL
中的distinct
类似
用法:
In [26]: data['company'].drop_duplicates()
Out[26]:
0 B
1 A
8 C
Name: company, dtype: object
.drop()
作用对象:Series
,DataFrame
主要用途:常用于删掉DataFrame
中的某些字段
主要参数:
- columns (single label or list-like)
指定要删掉的字段
用法:
# 删掉'gender'列
In [27]: data.drop(columns = ['gender'])
Out[27]:
number company salary
0 40 B 30
1 31 A 36
2 28 B 35
3 18 B 9
4 43 B 16
5 22 A 46
6 28 B 15
7 40 B 33
8 32 C 19
.isin()
作用对象:Series
,DataFrame
主要用途:常用于构建布尔索引,对DataFrame
的数据进行条件筛选
用法:
# 筛选出A公司和C公司的员工记录
In [29]: data.loc[data['company'].isin(['A','C'])]
Out[29]:
number company gender salary
1 31 A female 36
5 22 A male 46
8 32 C male 19
pd.cut()
主要用途:将连续变量离散化,比如将人的年龄划分为各个区间
主要参数:
- x (array-like)
需要进行离散化的一维数据 - bins (int, sequence of scalars, or IntervalIndex)
设置需要分成的区间,可以指定区间数量,也可以指定间断点 - labels (array or bool, optional)
设置区间的标签
用法:
# 把薪水分成5个区间
In [33]: pd.cut(data.salary,bins = 5)
Out[33]:
0 (23.8, 31.2]
1 (31.2, 38.6]
2 (31.2, 38.6]
3 (8.963, 16.4]
4 (8.963, 16.4]
5 (38.6, 46.0]
6 (8.963, 16.4]
7 (31.2, 38.6]
8 (16.4, 23.8]
Name: salary, dtype: category
Categories (5, interval[float64]): [(8.963, 16.4] < (16.4, 23.8] < (23.8, 31.2] < (31.2, 38.6] <(38.6, 46.0]]
# 自行指定间断点
In [32]: pd.cut(data.salary,bins = [0,10,20,30,40,50])
Out[32]:
0 (20, 30]
1 (30, 40]
2 (30, 40]
3 (0, 10]
4 (10, 20]
5 (40, 50]
6 (10, 20]
7 (30, 40]
8 (10, 20]
Name: salary, dtype: category
Categories (5, interval[int64]): [(0, 10] < (10, 20] < (20, 30] < (30, 40]