大数据统计笔记

# -*- coding: utf-8 -*-
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import json

s = pd.Series([1,3,5,np.nan,6,8]);
print(s);
s = pd.Series([1,3],["a","b"]);
print(s);

#循环查询日期范围内数据
dates = pd.date_range(‘20130101‘,periods=6);
print(dates);


df = pd.DataFrame(np.random.randn(6,4),index=dates,columns=list(‘ABCD‘))

print("LIMIT:============");
print(df.head(3));
print(df.tail(3));

print("反转XY显示 :============");
print(df.T);

#排序
#df.sort_index(axis=1, ascending=False)
#df.sort(columns=‘B‘)

 

########################选择器#############################
print("指定列 :============");
print(df[[‘A‘,‘B‘]]);


print("df[行范围,列范围]:============");
print(df[0:3]);

print("指定索引主键 :============");
print(df[‘20130103‘:‘20130104‘]);


print("布尔过滤 :============");
print(df[df.A>0.5]);

#print(df[0:3,0:1]);

########################距阵操作 #############################
print("距阵操作 :============");
print(df*2);
print(np.exp(df));
########################TABLE 元素#############################

#以二维数组显示
#df.values

#列描述
#df.columns

#索引
#df.index 
#对象类型
#df.dtypes

print("字典(JSON)转换距阵 :============");
sdata = {‘Ohio‘: 35000, ‘Texas‘: 71000, ‘Oregon‘: 16000, ‘Utah‘: 5000};
df = pd.Series(sdata);
print(df);

sdata = [{‘Ohio‘: 35000, ‘Texas‘: 71000, ‘Oregon‘: 16000, ‘Utah‘: 5000}];
df = pd.DataFrame(sdata);
df.columns = [‘Ohio‘, ‘Texas‘, ‘Oregon‘, ‘Utah‘];
print(df);
df = df.fillna(0) #将缺失值都替换为0

#入门
#http://pandas.pydata.org/pandas-docs/stable/10min.html
#http://pda.readthedocs.org/en/latest/chp5.html
#blog
#http://cloga.info/#wat_e_eb3d32d8-f59a-4a08-bf96-6f706d89c097_zss_

#大数据
#http://www.17bigdata.com/?cat=22
#优化
#http://1.aisensiy.sinaapp.com/2014/03/%E6%9C%80%E8%BF%91%E4%BD%BF%E7%94%A8-pandas-%E7%9A%84%E6%80%BB%E7%BB%93/
df.to_csv(‘E:\\py\\foo.csv‘)


######################JSON###########################

print("josn====================");
j = [{‘0‘:{"a":"a"},‘1‘:{"b":"b"}}];
elevations = json.dumps(j);
df = pd.read_json(elevations );
print(df);