Python之pandas模块

pandas

pandas is an open source, BSD-licensed library providing high-performance, easy-to-use data structures and data analysis tools for the Python programming language

一、安装

1、pandas

git clone git://github.com/pydata/pandas.git

2、Requirements

Cpython

https://pypi.python.org/pypi/Cython/

pytz

https://pypi.python.org

二、例子：

数据内容格式：

dstat命令输出的结果

"epoch","1m","5m","15m","read","writ","read","writ","usr","sys","idl","wai","hiq","siq","recv","send","used","buff","cach","free","util","util"
1384246698.290,2.730,1.190,0.940,4316699.658,3074071.976,283.904,131.427,1.172,0.464,96.302,1.858,0.022,0.182,0.0,0.0,6959415296.0,1933697024.0,19159023616.0
,22583123968.0,0.155,0.292
1384246699.290,2.730,1.190,0.940,4579328.0,880640.0,560.0,27.0,14.518,0.876,82.228,1.877,0.0,0.501,363556.0,2888961.0,6973390848.0,1933697024.0,19159031808.0
,22569140224.0,0.0,17.600
1384246700.290,2.730,1.190,0.940,3678208.0,516096.0,449.0,18.0,13.965,0.623,83.292,1.621,0.125,0.374,140745.0,1662141.0,6983143424.0,1933697024.0,19159306240
.0,22559113216.0,0.0,13.800
1384246701.290,2.510,1.170,0.930,4292608.0,667648.0,524.0,49.0,7.644,0.752,89.348,1.629,0.0,0.627,139858.0,1056176.0,6971334656.0,1933697024.0,19159306240.0,
22570921984.0,0.0,15.0
1384246702.290,2.510,1.170,0.930,4104192.0,684032.0,502.0,22.0,3.616,0.748,93.267,1.621,0.0,0.748,126114.0,1135695.0,6983221248.0,1933697024.0,19159592960.0,
22558748672.0,0.0,14.800

程序：

#!/usr/bin/env python2.7
#-*- coding:utf-8 -*-

import pandas as pd
import numpy as np

f = open("./ssd_dstat.txt")

index_list = []
flag = False

all_data_list=[]

for line in f.xreadlines():
    if "buff" in line:
        index_str_list = line.strip('
').split(',')

        index_list = []

        for i in index_str_list:

            column_name = i.strip('"')
            if column_name in index_list:                                                                                                                    
                column_name = column_name + '_1'

            index_list.append(column_name)

        print index_list

        flag = True
    elif flag:
        data_str_list = line.strip('
').split(',')

        data_list = []
        for i in data_str_list:
            data_list.append(float(i))

        all_data_list.append(data_list)


data = np.array(all_data_list)

df = pd.DataFrame(data,columns=index_list)

print df

以上是个初步的例子，还有待优化