Adam Wentz created AIRFLOW-1558:
-----------------------------------

             Summary: S3FileTransformOperator fails in Python 3 due to file mode
                 Key: AIRFLOW-1558
                 URL: https://issues.apache.org/jira/browse/AIRFLOW-1558
             Project: Apache Airflow
          Issue Type: Bug
          Components: operators
    Affects Versions: Airflow 1.8
         Environment: python3
            Reporter: Adam Wentz
            Priority: Minor


When running under python3 the S3FileTransformOperator fails with the following 
error:

```
[2017-09-01 18:44:54,440] {models.py:1427} ERROR - write() argument must be 
str, not bytes
[2017-09-01 18:44:54,443] {base_task_runner.py:95} INFO - Subtask: Traceback 
(most recent call last):
[2017-09-01 18:44:54,444] {base_task_runner.py:95} INFO - Subtask:   File 
"/usr/local/lib/python3.6/site-packages/airflow/models.py", line 1384, in run
[2017-09-01 18:44:54,445] {base_task_runner.py:95} INFO - Subtask:     result = 
task_copy.execute(context=context)
[2017-09-01 18:44:54,445] {base_task_runner.py:95} INFO - Subtask:   File 
"/usr/local/lib/python3.6/site-packages/airflow/operators/s3_file_transform_operator.py",
 line 87, in execute
[2017-09-01 18:44:54,446] {base_task_runner.py:95} INFO - Subtask:     
source_s3_key_object.get_contents_to_file(f_source)
[2017-09-01 18:44:54,446] {base_task_runner.py:95} INFO - Subtask:   File 
"/usr/local/airflow/.local/lib/python3.6/site-packages/boto/s3/key.py", line 
1662, in get_contents_to_file
[2017-09-01 18:44:54,447] {base_task_runner.py:95} INFO - Subtask:     
response_headers=response_headers)
[2017-09-01 18:44:54,447] {base_task_runner.py:95} INFO - Subtask:   File 
"/usr/local/airflow/.local/lib/python3.6/site-packages/boto/s3/key.py", line 
1494, in get_file
[2017-09-01 18:44:54,447] {base_task_runner.py:95} INFO - Subtask:     
query_args=None)
[2017-09-01 18:44:54,448] {base_task_runner.py:95} INFO - Subtask:   File 
"/usr/local/airflow/.local/lib/python3.6/site-packages/boto/s3/key.py", line 
1548, in _get_file_internal
[2017-09-01 18:44:54,448] {base_task_runner.py:95} INFO - Subtask:     
fp.write(bytes)
[2017-09-01 18:44:54,448] {base_task_runner.py:95} INFO - Subtask:   File 
"/usr/local/lib/python3.6/tempfile.py", line 483, in func_wrapper
[2017-09-01 18:44:54,449] {base_task_runner.py:95} INFO - Subtask:     return 
func(*args, **kwargs)
[2017-09-01 18:44:54,449] {base_task_runner.py:95} INFO - Subtask: TypeError: 
write() argument must be str, not bytes
[2017-09-01 18:44:54,450] {base_task_runner.py:95} INFO - Subtask: [2017-09-01 
18:44:54,443] {models.py:1451} INFO - Marking task as FAILED.
```
The solution is to open the `NamedTemporaryFile`s with mode `wb` rather than 
`w`. I have an incoming PR for this.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Reply via email to