主页 / 问题 / 1608670
Accepted
evgesha
Asked:2025-03-14 05:05:55 +0000 UTC2025-03-14 05:05:55 +0000 UTC

如何读取“损坏的” CSV？

  • 772

当读取 CSV 文件时，我遇到了数据可能损坏的问题。例如：

"Job ID,""Server"",""Agent"",""Subclient"",""Backup type"",""Start time"",""End time"",""Duration(mins)"",""Application size"",""Job status"",""Failure reason"""

要将文件中的行转换为 CSV，我使用：

csv_data = StringIO(data)

该字符串被解析为单个值，而不是标题。如何克服这个问题？

python

1 个回答

  1. Best Answer

    一般来说，错误会手动更正，或者请求重新上传。但在您的情况下，它看起来像一个重复的 CSV 文件条目。也就是说，您需要将这些行读取两次作为 CSV 数据。例如：

    from io import StringIO
import csv

raw_data = '''\
"Job ID,""Server"",""Start time"",""Failure reason"""
"100500,""Some server"",""01:02:03,123"",""Something happened, but I don't know what"""
"142542,""Some other server"",""11:22:23,123"",""Nothing happened, it just didn't work"""
'''

with StringIO(raw_data) as file:
    data = [*map(lambda x: next(csv.reader(StringIO(x[0]))), csv.reader(file))]

print(data)

# [['Job ID', 'Server', 'Start time', 'Failure reason'], 
#  ['100500', 'Some server', '01:02:03,123', "Something happened, but I don't know what"], 
#  ['142542', 'Some other server', '11:22:23,123', "Nothing happened, it just didn't work"]]

    如果您有数据上传管道，最好检查 CSV 格式中是否有重复的数据条目。

    • 2

相关问题