有关csv格式的数据转换成libsvm格式的数据

# -*- coding: utf-8 -*-
"""
Created on Sat May 13 21:50:03 2017

@author: Administrator
"""
import pandas as pd
import time
def df2ffm(df, fp):
'''
Convert pandas.DataFrame to data format that libffm can directly use

@Args:
df: pandas.DataFrame to be converted
fp: save libffm format data to fp<filepath>
'''
now = time.time()
print('Format Converting begin in time:...',now)
columns = df.columns.values
d = len(columns)
feature_index = [i for i in range(d)]
field_index = [0]*d
field = []
for col in columns:
field.append(col.split('_')[0])
index = -1
for i in range(d):
if i==0 or field[i]!=field[i-1]:
index+=1
field_index[i] = index

with open(fp, 'w') as f:
for row in df.values:
line =str(int(row[0]))
for i in range(1, len(row)):
if row[i]!=0:
line += " %d:%d" % (feature_index[i], row[i])
line+='\n'
f.write(line)
print('finish convert,the cost time is ',time.time()-now)
print('[Done]')
print()

def main():
df = pd.read_csv(r'E:\tencent\input\train1.csv')
df = df.fillna(0)
fp = r'E:\tencent\output\train.csv'
df2ffm(df,fp)

if __name__ == '__main__':
main()