CSV文件的读取，TensorFlow和pandas

csv文件的读取，有两种方法：调用pandas库函数或者直接用TensorFlow读取，

1、调用pandas

data.csv是自己随便搞的一个数据文件，数据样例和读取代码如下：

import tensorflow as tf
import pandas as pd

def pd_read_csv():
 
    data_frame = pd.read_csv("data.csv", sep=",")
    data1 = []
    data2 = []
    data3 = []
    data4 = []
   
    for index in data_frame.index:
        data_row = data_frame.loc[index]
        data1.append(data_row["one"])
        data2.append(data_row["two"])
        data3.append(data_row["three"])
        data4.append(data_row["label"])
    return (data1,data2,data3,data4)

data1, data2, data3, data4 = pd_read_csv()

数据的第一行都是有特征的，比如one、two.......，当然也可以是纯数字，只需要修改