使用pandas读取csv文件的指定列方法

蓝鲸

发布时间：2026-10-13| 3人浏览过

原创

本文介绍如何利用pandas库中的usecols参数，在读取CSV文件时仅提取指定的列数据。通过对比全量读取与指定列读取的输出结果，展示了该方法在清理无效列、简化数据展示及提升终端界面清爽度方面的实际应用价值。

在成功实现读取CSV文件前几行数据后，自然联想到是否也能实现仅读取前几列的功能。经过多次尝试，最终找到了一种可行的解决方案。

之所以需要读取前几列，是因为手头的一份CSV文件包含若干无实际数据的尾部列，但这些列始终存在。原始数据格式如下：

GreydeMac-mini:chapter06 greyzhang$ cat data.csv

1,name_01,coment_01,,,,
2,name_02,coment_02,,,,
3,name_03,coment_03,,,,
4,name_04,coment_04,,,,
5,name_05,coment_05,,,,
6,name_06,coment_06,,,,
7,name_07,coment_07,,,,
8,name_08,coment_08,,,,
9,name_09,coment_09,,,,
10,name_10,coment_10,,,,
11,name_11,coment_11,,,,
12,name_12,coment_12,,,,
13,name_13,coment_13,,,,
14,name_14,coment_14,,,,
15,name_15,coment_15,,,,
16,name_16,coment_16,,,,
17,name_17,coment_17,,,,
18,name_18,coment_18,,,,
19,name_19,coment_19,,,,
20,name_20,coment_20,,,,
21,name_21,coment_21,,,,

若使用pandas直接读取全部数据，打印输出时将呈现如下结果：

In [41]: data = pd.read_csv('data.csv')

In [42]: data
Out[42]: 
  1 name_01 coment_01 Unnamed: 3 Unnamed: 4 Unnamed: 5 Unnamed: 6
0 2 name_02 coment_02   NaN   NaN   NaN   NaN
1 3 name_03 coment_03   NaN   NaN   NaN   NaN
2 4 name_04 coment_04   NaN   NaN   NaN   NaN
3 5 name_05 coment_05   NaN   NaN   NaN   NaN
4 6 name_06 coment_06   NaN   NaN   NaN   NaN
5 7 name_07 coment_07   NaN   NaN   NaN   NaN
6 8 name_08 coment_08   NaN   NaN   NaN   NaN
7 9 name_09 coment_09   NaN   NaN   NaN   NaN
8 10 name_10 coment_10   NaN   NaN   NaN   NaN
9 11 name_11 coment_11   NaN   NaN   NaN   NaN
10 12 name_12 coment_12   NaN   NaN   NaN   NaN
11 13 name_13 coment_13   NaN   NaN   NaN   NaN
12 14 name_14 coment_14   NaN   NaN   NaN   NaN
13 15 name_15 coment_15   NaN   NaN   NaN   NaN
14 16 name_16 coment_16   NaN   NaN   NaN   NaN
15 17 name_17 coment_17   NaN   NaN   NaN   NaN
16 18 name_18 coment_18   NaN   NaN   NaN   NaN
17 19 name_19 coment_19   NaN   NaN   NaN   NaN
18 20 name_20 coment_20   NaN   NaN   NaN   NaN
19 21 name_21 coment_21   NaN   NaN   NaN   NaN

虽然在学习阶段这并非障碍，但长期在命令行终端操作时，更倾向于清爽简洁的输出风格。利用read_csv函数的usecols参数，可在一定程度上减少这种视觉混乱。

In [45]: data = pd.read_csv('data.csv',usecols=[0,1,2,3])

In [46]: data
Out[46]: 
  1 name_01 coment_01 Unnamed: 3
0 2 name_02 coment_02   NaN
1 3 name_03 coment_03   NaN
2 4 name_04 coment_04   NaN
3 5 name_05 coment_05   NaN
4 6 name_06 coment_06   NaN
5 7 name_07 coment_07   NaN
6 8 name_08 coment_08   NaN
7 9 name_09 coment_09   NaN
8 10 name_10 coment_10   NaN
9 11 name_11 coment_11   NaN
10 12 name_12 coment_12   NaN
11 13 name_13 coment_13   NaN
12 14 name_14 coment_14   NaN
13 15 name_15 coment_15   NaN
14 16 name_16 coment_16   NaN
15 17 name_17 coment_17   NaN
16 18 name_18 coment_18   NaN
17 19 name_19 coment_19   NaN
18 20 name_20 coment_20   NaN
19 21 name_21 coment_21   NaN

为清晰展示数据“边界”，示例中保留了第一列无效数据。在实际应用中，若希望去除结果中最后一列信息，只需在参数列表中移除对应列号即可。

In [47]: data = pd.read_csv('data.csv',usecols=[0,1,2])

In [48]: data
Out[48]: 
  1 name_01 coment_01
0 2 name_02 coment_02
1 3 name_03 coment_03
2 4 name_04 coment_04
3 5 name_05 coment_05
4 6 name_06 coment_06
5 7 name_07 coment_07
6 8 name_08 coment_08
7 9 name_09 coment_09
8 10 name_10 coment_10
9 11 name_11 coment_11
10 12 name_12 coment_12
11 13 name_13 coment_13
12 14 name_14 coment_14
13 15 name_15 coment_15
14 16 name_16 coment_16
15 17 name_17 coment_17
16 18 name_18 coment_18
17 19 name_19 coment_19
18 20 name_20 coment_20
19 21 name_21 coment_21

以上即为使用pandas读取CSV文件指定列的完整方法，希望对大家有所帮助，欢迎参考与支持。