1.describe()：查看每一列的描述性统计量

# 导包
import numpy as np
import pandas as pd

df = pd.DataFrame(data=np.random.randint(0,10,size=(5,3)),
                  index=list("ABCDE"),
                  columns=["Python","NumPy","Pandas"]
                 )
df

df.describe()

	Python	NumPy	Pandas
count	5.000000	5.000000	5.000000
mean	5.600000	2.800000	5.400000
std	2.073644	2.167948	2.408319
min	4.000000	0.000000	2.000000
25%	4.000000	1.000000	4.000000
50%	5.000000	4.000000	6.000000
75%	6.000000	4.000000	7.000000
max	9.000000	5.000000	8.000000

# 自定义百分数
df.describe([0.01,0.3,0.4,0.9,0.99])

	Python	NumPy	Pandas
count	5.000000	5.000000	5.000000
mean	5.600000	2.800000	5.400000
std	2.073644	2.167948	2.408319
min	4.000000	0.000000	2.000000
1%	4.000000	0.040000	2.080000
30%	4.200000	1.600000	4.400000
40%	4.600000	2.800000	5.200000
50%	5.000000	4.000000	6.000000
90%	7.800000	4.600000	7.600000
99%	8.880000	4.960000	7.960000
max	9.000000	5.000000	8.000000

# 行列转置
df.describe([0.01,0.3,0.4,0.9,0.99]).T

	count	mean	std	min	1%	30%	40%	50%	90%	99%	max
Python	5.0	5.6	2.073644	4.0	4.00	4.2	4.6	5.0	7.8	8.88	9.0
NumPy	5.0	2.8	2.167948	0.0	0.04	1.6	2.8	4.0	4.6	4.96	5.0
Pandas	5.0	5.4	2.408319	2.0	2.08	4.4	5.2	6.0	7.6	7.96	8.0

2.df.std()：可以求得DataFrame对象每一列的标准差

df.std()

Python    2.073644
NumPy     2.167948
Pandas    2.408319
dtype: float64

3.df.drop()：删除特定索引

df2 = df.copy()
df2

	Python	NumPy	Pandas
A	9	0	8
B	5	1	2
C	6	5	7
D	4	4	6
E	4	4	4

# 默认删除行
df2.drop("A")

	Python	NumPy	Pandas
B	5	1	2
C	6	5	7
D	4	4	6
E	4	4	4

df2.drop(index="A")

	Python	NumPy	Pandas
B	5	1	2
C	6	5	7
D	4	4	6
E	4	4	4

# 删除列
df2.drop("Python",axis=1)

	NumPy	Pandas
A	0	8
B	1	2
C	5	7
D	4	6
E	4	4

df2.drop(columns="Python")

	NumPy	Pandas
A	0	8
B	1	2
C	5	7
D	4	6
E	4	4

# 删除多列或多行
df2.drop(columns=["NumPy","Python"])

	Pandas
A	8
B	2
C	7
D	6
E	4

df2.drop(index=["A","B"])

	Python	NumPy	Pandas
C	6	5	7
D	4	4	6
E	4	4	4

# inplace修改原数据
df2.drop(index=["A","B"],inplace=True)
df2

	Python	NumPy	Pandas
C	6	5	7
D	4	4	6
E	4	4	4

4.unique()：唯一，去重(只能用于Series一维数组)

# DataFrame没有unique属性，Series调用unique
df["Python"].unique()

array([9, 5, 6, 4])

5.df.query：按条件查询

# 找到Python列中等于6的所有行
df.query("Python == 6")

	Python	NumPy	Pandas
C	6	5	7

df.query("Python > 6")

	Python	NumPy	Pandas
A	9	0	8

df.query("Python < 6")

	Python	NumPy	Pandas
B	5	1	2
D	4	4	6
E	4	4	4

# and,&
df.query("Python > 4 and NumPy == 5")

	Python	NumPy	Pandas
C	6	5	7

df.query("Python > 4 & NumPy == 5")

	Python	NumPy	Pandas
C	6	5	7

# or,|
df.query("Python > 6 or NumPy == 8")

	Python	NumPy	Pandas
A	9	0	8

df.query("Python > 6 | NumPy == 8")

	Python	NumPy	Pandas
A	9	0	8

# in（成员运算符）
df.query("Python in [5,6,9]")

	Python	NumPy	Pandas
A	9	0	8
B	5	1	2
C	6	5	7

# 使用变量@符号
n = 6
df.query("Python == @n")

	Python	NumPy	Pandas
C	6	5	7

m = [5,6,9]
df.query("Python in @m")

	Python	NumPy	Pandas
A	9	0	8
B	5	1	2
C	6	5	7

6.df.sort_values()：根据值排序

# sort_values：默认按照列名排序，默认升序（常用）
df.sort_values("Python")

	Python	NumPy	Pandas
D	4	4	6
E	4	4	4
B	5	1	2
C	6	5	7
A	9	0	8

# ascending：是否升序，默认是True
df.sort_values("Python",ascending=False)

	Python	NumPy	Pandas
A	9	0	8
C	6	5	7
B	5	1	2
D	4	4	6
E	4	4	4

# 根据行索引名排序，会把列进行排序（不常用）
df.sort_values("B",axis=1)

	NumPy	Pandas	Python
A	0	8	9
B	1	2	5
C	5	7	6
D	4	6	4
E	4	4	4

7.df.sort_index()：根据索引排序

# 按照索引名排序，默认是对行索引进行排序，默认是升序
df.sort_index(ascending=False)

	Python	NumPy	Pandas
E	4	4	4
D	4	4	6
C	6	5	7
B	5	1	2
A	9	0	8

# 按照列索引排序
df.sort_index(ascending=False,axis=1)

	Python	Pandas	NumPy
A	9	8	0
B	5	2	1
C	6	7	5
D	4	6	4
E	4	4	4

8.df.info()：查看数据信息

# info：常用
df.info()

<class 'pandas.core.frame.DataFrame'>
Index: 5 entries, A to E
Data columns (total 3 columns):
 #   Column  Non-Null Count  Dtype
---  ------  --------------  -----
 0   Python  5 non-null      int32
 1   NumPy   5 non-null      int32
 2   Pandas  5 non-null      int32
dtypes: int32(3)
memory usage: 272.0+ bytes

9.练习

• 新建一个形状为10000*3的标准正态分布的DataFrame(np.random.randn)，去除掉所有满足以下情况的行：其中任一元素绝对值大约3陪标准差

df = pd.DataFrame(np.random.randn(10000,3))
df

	0	1	2
0	0.786386	-0.204965	-0.152465
1	0.400526	1.447733	0.310461
2	-0.363709	-0.989258	-1.093327
3	-2.856978	2.336645	1.474821
4	-0.847757	1.141278	-0.230877
...	...	...	...
9995	0.331052	0.263059	0.469468
9996	1.172189	-1.380337	0.648793
9997	-0.544049	-0.509627	-0.224698
9998	-0.034967	-0.085575	-0.687314
9999	0.007202	-0.069250	-0.803754

10000 rows × 3 columns

# 过滤掉 大于3陪标准差的行
# 标准差 df.std()，绝对值 df.abs()
# cond：找到每一个元素是否大于3陪标准差
cond = df.abs() > df.std()*3
cond

	0	1	2
0	False	False	False
1	False	False	False
2	False	False	False
3	False	False	False
4	False	False	False
...	...	...	...
9995	False	False	False
9996	False	False	False
9997	False	False	False
9998	False	False	False
9999	False	False	False

10000 rows × 3 columns

# 找到存在大于3陪标准差的行
cond2 = cond.any(axis=1)
cond2

0       False
1       False
2       False
3       False
4       False
        ...  
9995    False
9996    False
9997    False
9998    False
9999    False
Length: 10000, dtype: bool

# bool值索引，过滤异常值（大于3陪标准差）
df.loc[~cond2]

	0	1	2
0	0.786386	-0.204965	-0.152465
1	0.400526	1.447733	0.310461
2	-0.363709	-0.989258	-1.093327
3	-2.856978	2.336645	1.474821
4	-0.847757	1.141278	-0.230877
...	...	...	...
9995	0.331052	0.263059	0.469468
9996	1.172189	-1.380337	0.648793
9997	-0.544049	-0.509627	-0.224698
9998	-0.034967	-0.085575	-0.687314
9999	0.007202	-0.069250	-0.803754

9904 rows × 3 columns