Pandas 介紹

pandas是python的一個數(shù)據(jù)分析庫，主要提供兩種主要的資料結(jié)構(gòu)，Series與DataFrame。Series是用來處理時間 順序相關(guān)資料，DataFrame則是用來處理結(jié)構(gòu)化的資料（二維的數(shù)據(jù)資料）

安裝Pandas

pip install pandas

Pandas讀取不同格式的資料

讀取CSV檔案

import pandas as pd
df = pd.read_csv('file.csv')
print(df)

讀取HTML檔案

import pandas as pd
df = pd.read_html('http://www.itdecent.cn/u/e635858eda0b')
print(df)

Pandas提供的資料結(jié)構(gòu)

· Series：處理時間序列的相關(guān)資料，主要是創(chuàng)建一維list。

·DataFrame：處理結(jié)構(gòu)化的資料，有索引和標簽的二維資料集。

·Panel：處理三維數(shù)據(jù)。

1.series

數(shù)據(jù)類型是array

 import pandas as pd
 list = ['python', 'ruby', 'c', 'c++']
 select = pd.Series(list)
 print (select)
輸出：
0    python
1      ruby
2         c
3       c++
dtype: object

數(shù)據(jù)類型是Dictionary

import pandas as pd
dict = {'key1': '1', 'key2': '2', 'key3': '3'}
select = pd.Series(dict, index = dict.keys())
輸出：
print select
key3    3
key2    2
key1    1
dtype: object
print (select[0])
3
print select[2]
1
print select['key3']
3
print select[[2]]
key1    1
dtype: object
print (select[[0,2,1]])
key3    3
key1    1
key2    2
dtype: object

數(shù)據(jù)類型是單一數(shù)據(jù)

import pandas as pd
string = 'henry'
 select = pd.Series (string, index = range(3))
 print (select)
輸出：
0    henry
1    henry
2    henry

切片選擇

 print (select[1:])
1    henry
2    henry

2.DataFrame

2.1建立DataFrame

可以用DDictionary或Array來創(chuàng)建，也可以用外部資料讀取后創(chuàng)建。

Dictionary

import pandas as pd
groups = ['Movies', 'Sports', 'Conding', 'Fishing', 'Dancing']
num = [12, 5, 18, 99, 88]
dict = {'groups': groups, 'num': num}
df = pd.DataFrame(dict)
print (df)
輸出：
    groups  num
0   Movies   12
1   Sports    5
2  Conding   18
3  Fishing   99
4  Dancing   88

Array

 array = [['Movies',12], ['Sports', 5], ['Conding', 18], ['Fishing', 99], ['Dancing', 88]]
df = pd.DataFrame(arr, colums = ['name', 'num'])
df = pd.DataFrame(array, columns = ['name', 'num'])
print df
輸出：
      name  num
0   Movies   12
1   Sports    5
2  Conding   18
3  Fishing   99
4  Dancing   88

2.2DataFrame的操作

DataFrame的方法

.shape 返回行數(shù)和列數(shù)

.describe() 返回描述性統(tǒng)計

.head()

.tail()

.columns

.index

.info()

import pandas as pd
groups = ['Movies', 'Sports', 'Conding', 'Fishing', 'Dancing']
num = [12, 5, 18, 99, 88]
dict = {'groups': groups, 'num': num}
df = pd.DataFrame(dict)
print df.shape
(5, 2)
print df.describe()
             num
count   5.000000
mean   44.400000
std    45.224993
min     5.000000
25%    12.000000
50%    18.000000
75%    88.000000
max    99.000000
print df.head()
    groups  num
0   Movies   12
1   Sports    5
2  Conding   18
3  Fishing   99
4  Dancing   88
print df.columns
Index([u'groups', u'num'], dtype='object')
print df.index
RangeIndex(start=0, stop=5, step=1)
print df.info
<bound method DataFrame.info of     groups  num
0   Movies   12
1   Sports    5
2  Conding   18
3  Fishing   99
4  Dancing   88>
print df.tail(3)
    groups  num
2  Conding   18
3  Fishing   99
4  Dancing   88