Help Center/ ModelArts/ Troubleshooting/ Training Jobs/ In-Cloud Migration Adaptation Issues/ Error Occurred When Pandas Reads Data from an OBS File If MoXing Is Used to Adapt to an OBS Path
Updated on 2022-12-08 GMT+08:00

Error Occurred When Pandas Reads Data from an OBS File If MoXing Is Used to Adapt to an OBS Path

Symptom

If MoXing is used to adapt to an OBS path, an error occurs when pandas of a later version reads data from an OBS file.
1. 'can't decode byte xxx in position xxx'
2. 'OSError:File isn't open for writing'

Possible Causes

MoXing does not support Pandas of a later version.

Solution

  1. After the OBS path is adapted, change the file access mode from r to rb and change the _write_check_passed value in mox.file.File to True, as shown in the following is sample code:
    import pandas as pd
    import moxing as mox
    
    mox.file.shift('os', 'mox')  # Replace the open operation of the operating system with the operation for adapting the mox.file.File to the OBS path.
    
    param = {'encoding': 'utf-8'}
    path = 'xxx.csv'
    with open(path, 'rb') as f:
        f._wirte_check_passed = True
        df = pd.read_csv(ff, **param)
  2. Use the local PyCharm to remotely access notebook for debugging.

Summary and Suggestions

Before creating a training job, use the ModelArts development environment to debug the training code to maximally eliminate errors in code migration.