用python获取.csv文件某一列或者某些列

1.把三个csv文件中的feature值整合到一个文件中，同时添加相应的label。

1.  
   # -*-coding:utf-8 -*-
2.  
   import csv;
3.  
   label1 = '1'
4.  
   label2 = '2'
5.  
   label3 = '3'
6.  
   a = "feature1,feature2,feature3,feature4,feature5,feature6,feature7,feature8,feature9,feature10,label" "\n"
7.  
   with open("./dataset/dataTime2.csv", 'a') as rfile:
8.  
   rfile.writelines(a)
9.  
   with open("./dataset/f02.csv", 'rb') as file:
10.  
    a = file.readline().strip()
11.  
    while a:
12.  
    a = a ',' label1 "\n"
13.  
    #a = label1 ',' a "\n"
14.  
    with open("./dataset/dataTime2.csv", 'a') as rfile:
15.  
    rfile.writelines(a)
16.  
    a = file.readline().strip()
17.  
    with open("./dataset/g03.csv", 'rb') as file:
18.  
    a = file.readline().strip()
19.  
    while a:
20.  
    a = a ',' label2 "\n"
21.  
    #a = label2 ',' a "\n"
22.  
    with open("./dataset/dataTime2.csv", 'a') as rfile:
23.  
    rfile.writelines(a)
24.  
    a = file.readline().strip()
25.  
    with open("./dataset/normal05.csv", 'rb') as file:
26.  
    a = file.readline().strip()
27.  
    while a:
28.  
    a = a ',' label3 "\n"
29.  
    #a = label3 ',' a "\n"
30.  
    with open("./dataset/dataTime2.csv", 'a') as rfile:
31.  
    rfile.writelines(a)
32.  
    a = file.readline().strip()

2.获取csv文件中某一列，下面可以获得label为表头的列中对应的所有数值。

1.  
   filename = "./dataset/dataTime2.csv"
2.  
   list1 = []
3.  
   with open(filename, 'r') as file:
4.  
   reader = csv.DictReader(file)
5.  
   column = [row['label'] for row in reader]

3.获取csv文件中某些列，下面可以获得除label表头的对应列之外所有数值。

1.  
   import pandas as pd
2.  
   odata = pd.read_csv(filename)
3.  
   y = odata['label']
4.  
   x = odata.drop(['label'], axis=1) #除去label列之外的所有feature值

4.也可以处理成list[np.array]形式的数据。

1.  
   filename = "./dataset/dataTime2.csv"
2.  
   list1 = []
3.  
   with open(filename, 'r') as file:
4.  
   a = file.readline()
5.  
   while a:
6.  
   c = np.array(a.strip("\n").split(","))
7.  
   list1.append(c)

5.也可以处理成tensor格式数据集

1.  
   # -*-coding:utf-8 -*-
2.  
   import tensorflow as tf
3.  
   # 读取的时候需要跳过第一行
4.  
   filename = tf.train.string_input_producer(["./dataset/dataTime.csv"])
5.  
   reader = tf.TextLineReader(skip_header_lines=1)
6.  
   key, value = reader.read(filename)
7.  
   record_defaults = [[1.], [1.], [1.], [1.], [1.], [1.], [1.], [1.], [1.], [1.], tf.constant([], dtype=tf.int32)]
8.  
   col1, col2, col3, col4, col5, col6, col7, col8, col9, col10, col11= tf.decode_csv(
9.  
   value, record_defaults=record_defaults)
10.  
    features = tf.stack([col1, col2, col3, col4, col5, col6, col7, col8, col9, col10])
11.  
    with tf.Session() as sess:
12.  
    # Start populating the filename queue.
13.  
    coord = tf.train.Coordinator()
14.  
    threads = tf.train.start_queue_runners(coord=coord)
15.  
    trainx = []
16.  
    trainy = []
17.  
    for i in range(81000):
18.  
    # Retrieve a single instance:
19.  
    example, label = sess.run([features, col11])
20.  
    trainx.append(example)
21.  
    trainy.append(label)
22.  
    coord.request_stop()
23.  
    coord.join(threads)
24.  
    #最后长度是81000，trainx是10个特征

参考资料：http://t.csdn.cn/HFTPy

这篇好文章是转载于：学新通技术网