pands 编码知识

一，pandas功能

    1，基于numpy , 分析结构化数据。

二，常用基础知识编码练习

    包括数据类型，数据操作，比如索引，分片 ，分组聚合 ，排序 过滤等等数分常见操作代码

# coding=utf-8

import pandas as pd
import numpy as np
import sklearn as sk
import matplotlib as mp
'''
data analisis structured data  baseed on numpy
data file : csv json sql MS excel
manipulation: map  reduce group aggregation sorting query fillter
join  selection  ETL   feature enginnering
same as spark flink Hsql 
data structure: 
  Series: one column ,any type , 
  DataFrame: table , multipul column 

'''
'''
print pd.__version__
mydataset = {
    'sites' : ["Google", "Runoob", "Wiki"],
    'number': [1, 2, 3]
}
list = [1, 2, 3]
myvar = pd.DataFrame(mydataset)
series = pd.Series(list)
print series[1]

series1 = pd.Series(list, index = ["x", "y", "z"])
print series1

data = [['Google', 10],['Runoob', 12],['Wiki', 13]] # nested list
df = pd.DataFrame(data, columns = ['Site', 'Age'], dtype= float)
print (df)

data_1 = {'Site':['Google', 'Runoob', 'Wiki'],'Age':[10,12,13]}
df_1 = pd.DataFrame(data_1)
# pd.DataFrame(data_1, index=["a","b"])
print (df_1.loc[1])   #row
# df_1.loc[0,1]

# read file
df = pd.read_csv("D:\\vbox-share\\better-training-for-safer-foods-april-2014-to-march-2015.csv").head(10)
print (df)
print (df.info())

print np.zeros((2,3))
print np.random.rand(10)


a = np.arange(6).reshape(3, 2)
print a

'''

data = np.array(['a','b','c','d','e'])
ser = pd.Series(data,index=[3,2,6,8,9]) #索引
print (ser[9])

df = pd.read_csv("C:\\Users\\86187\\Downloads\\nba.csv") #读取文件
ser = pd.Series(df['Name'])
data = ser.head(10)  //切片
# print (data[3:6])
print (data.loc[3:6])

# binary operations like add sub 
sub_d1 = pd.Series([5,2,4,6], index=['a','b','c','d'])
sub_d2 = pd.Series([1,2,6,7], index=['a','b','d','e'])
print  sub_d1.sub(sub_d2,fill_value=0)

# add sub mul div sum prod


# dataframe  indexing 抽取数据 ， selection and prejection
df_obj = pd.DataFrame(np.random.rand(5,4),columns=['a','b','c','d'])
print (df_obj.head())
print (df_obj['a'])  #df_obj['a','c']
print (df_obj.loc[0:2,'a']) #df_obj.iloc[0:2, 0]

#对齐运算
 # Series 对齐  dataframe  2个df ， 元素对元素
align_df1= pd.DataFrame(np.ones((2,2)),columns=['a','b'])
align_df2= pd.DataFrame(np.ones((3,3)),columns=['a','b','c'])
print (align_df1 + align_df2)
# print (align_df1.add(align_df2,fill_value = -1))  未对齐填充值

# function computaion   单个df， 对每个元素，行 列方向使用函数计算
func_df = pd.DataFrame(np.random.rand(5,4) - 1)
print (np.abs(func_df))  # np 函数
print (func_df.apply(lambda x: x.max()))  # 默认axis=0，方向是列

f2 = lambda x: '%.2f' % x
print (func_df.applymap(f2)) # each element

# sorting
sort_s1 = pd.Series(range(10,15), index=np.random.randint(5, size=5))
print (sort_s1)
print (sort_s1.sort_index()) # ascending default, by index column
# df.sort_index(axis=1,ascending=False)  df考虑轴方向，1 列方向排序。
print (func_df.sort_values(by = 0,ascending=False)) # 0 列名，的值来排序。

# deal with missing value
missing_value_df = pd.DataFrame([np.random.rand(3),[1.,2.,np.nan]
                                [np.nan,4.,np.nan],[1.,2.,3.]])

以上执行后结果：

以上为本人练习pandas功能，并做了笔记。