python学习笔记（四）：pandas基础

pandas 基础

serise

import pandas as pd
from pandas import Series, DataFrame
obj = Series([4, -7, 5, 3])
obj

0 4
1 -7
2 5
3 3
dtype: int64

obj.values

array([ 4, -7, 5, 3], dtype=int64)

obj.index

RangeIndex(start=0, stop=4, step=1)

obj[[1,3]]
# 跳着选取数据

1 -7
3 3
dtype: int64

obj[1:3]

1 -7
2 5
dtype: int64

pd.isnull(obj)

0 False
1 False
2 False
3 False
dtype: bool

• reindex可以用来插值

obj.reindex(range(5), method = 'ffill')

0 4
1 -7
2 5
3 3
4 3
dtype: int64

• 标签切片是闭区间的

dataframe

data = {'state': ['asd','qwe','sdf','ert'],
       'year': [2000, 2001, 2002, 2003],
       'pop': [1.5,1.7,3.6,2.4]}
data = DataFrame(data)
data

	pop	state	year
0	1.5	asd	2000
1	1.7	qwe	2001
2	3.6	sdf	2002
3	2.4	ert	2003

data.year
# 比r里提取列要方便点

0 2000
1 2001
2 2002
3 2003
Name: year, dtype: int64

data['debt'] = range(4)
data

	pop	state	year	debt
0	1.5	asd	2000	0
1	1.7	qwe	2001	1
2	3.6	sdf	2002	2
3	2.4	ert	2003	3

• index是不能修改的

a = data.index
a[1] = 6

---------------------------------------------------------------------------

TypeError Traceback (most recent call last)

in ()
1 a = data.index
----> 2 a[1] = 6

F:\Anaconda\lib\site-packages\pandas\core\indexes\base.py in setitem(self, key, value)
1668
1669 def setitem(self, key, value):
-> 1670 raise TypeError("Index does not support mutable operations")
1671
1672 def getitem(self, key):

TypeError: Index does not support mutable operations

data.columns

Index(['pop', 'state', 'year', 'debt'], dtype='object')

• .ix标签索引功能，输入行和列
• 不加.ix只能选取其中的某列或某行，不能列与行同时选取

data[:3]

	pop	state	year	debt
0	1.5	asd	2000	0
1	1.7	qwe	2001	1
2	3.6	sdf	2002	2

data.ix[:,:3]

	pop	state	year
0	1.5	asd	2000
1	1.7	qwe	2001
2	3.6	sdf	2002
3	2.4	ert	2003

• 删除某列用drop，axis = 0表示行，1表示列
• 删除后原数据不变

data.drop(0,axis=0)

	pop	state	year	debt
1	1.7	qwe	2001	1
2	3.6	sdf	2002	2
3	2.4	ert	2003	3

data.drop('year', axis=1)

	pop	state	debt
0	1.5	asd	0
1	1.7	qwe	1
2	3.6	sdf	2
3	2.4	ert	3

data

	pop	state	year	debt
0	1.5	asd	2000	0
1	1.7	qwe	2001	1
2	3.6	sdf	2002	2
3	2.4	ert	2003	3

import numpy as np
df = DataFrame(np.arange(9).reshape(3, 3))
df

	0	1	2
0	0	1	2
1	3	4	5
2	6	7	8

• applymap()可以对dataframe每一个元素运用函数
• apply()可以对每一维数组运用函数

df.applymap(lambda x: '%.2f' % x)

	0	1	2
0	0.00	1.00	2.00
1	3.00	4.00	5.00
2	6.00	7.00	8.00

data.sort_values(by='pop')
# 对某一列排序

	pop	state	year	debt
0	1.5	asd	2000	0
1	1.7	qwe	2001	1
3	2.4	ert	2003	3
2	3.6	sdf	2002	2

data.describe()

	pop	year	debt
count	4.000000	4.000000	4.000000
mean	2.300000	2001.500000	1.500000
std	0.948683	1.290994	1.290994
min	1.500000	2000.000000	0.000000
25%	1.650000	2000.750000	0.750000
50%	2.050000	2001.500000	1.500000
75%	2.700000	2002.250000	2.250000
max	3.600000	2003.000000	3.000000

df.isin([1])

	0	1	2
0	False	True	False
1	False	False	False
2	False	False	False

• None、NaN会被当作NA处理
• df.shape不加括号相当于dim()

df.shape

(3, 3)

• dropna删除缺失值

df.ix[:1, :1] = None
df

	0	1	2
0	NaN	NaN	2
1	NaN	NaN	5
2	6.0	7.0	8

• 填充缺失值可以调用字典，不同行添加不同值

df.fillna({0:11, 1:22})

	0	1	2
0	11.0	22.0	2
1	11.0	22.0	5
2	6.0	7.0	8

df

	0	1	2
0	NaN	NaN	2
1	NaN	NaN	5
2	6.0	7.0	8

df.fillna({0:11, 1:22}, inplace=True)

	0	1	2
0	11.0	22.0	2
1	11.0	22.0	5
2	6.0	7.0	8

df

	0	1	2
0	11.0	22.0	2
1	11.0	22.0	5
2	6.0	7.0	8

• inplace修改对象不产生副本