gemelen opened a new issue, #26952:
URL: https://github.com/apache/airflow/issues/26952
### Apache Airflow Provider(s)
google
### Versions of Apache Airflow Providers
apache-airflow-providers-google==2022.9.6+composer
apache-airflow-providers-google==8.1.0
### Apache Airflow version
2.2.3 | 2.3.3
### Operating System
Managed
### Deployment
Composer
### Deployment details
apache-airflow-providers-google==2022.9.6+composer (provided by
composer-2.0.28-airflow-2.2.5)
apache-airflow-providers-google==8.1.0 (provided by
composer-2.0.28-airflow-2.3.3)
Cluster details are identical between versions.
Data that DAG/jobs are processing are identical between versions.
### What happened
My original DAG consists of a list of BigQueryInsertJobOperator, each of
them is essentially running `select * from <source dataset>.<source table>
where <field> = <value from list>` and inserting result records into
<destination dataset>.<destination table>, where `<destination dataset> =
<source table>'s name` and `<destination table> = <source table>_<field value>`.
Operator is configured in a next way:
```
bigquery.BigQueryInsertJobOperator(
...
configuration = {
'query': {
'useLegacySql': False,
'priority': 'BATCH',
'createDisposition': 'CREATE_IF_NEEDED',
'writeDisposition': 'WRITE_TRUNCATE',
'destinationTable': {
'projectId': GCP_PROJECT,
'datasetId': table,
'tableId': f'{table}_{business_id}'
},
'defaultDataset': {
'projectId': GCP_PROJECT,
'datasetId': DATASET,
},
'query': f'select * from {table} where business_id =
{business_id};'
},
}
)
```
In an environment of Composer `composer-2.0.28-airflow-2.2.5` (ie with a
provided lib of version `2022.9.6+composer`) it works as intended (job is
inserted and executed).
In an environment of Composer `composer-2.0.28-airflow-2.3.3` (ie with a
provided lib of version `8.1.0`) each job is inserted but fails with an error
`"Dataset was not found in location"` (for the same set of datasets and tables).
The same job inserted via BigQuery REST API executed successfully too.
### What you think should happen instead
As described above, job should succeed.
### How to reproduce
_No response_
### Anything else
Error occurs for every run, and most likely is due to the changes in
dataset/table reference, so in a latter version (from versions above) its
reference incorrectly (ie using wrong separator, for example).
As a result, this issue prevents from upgrade to the version of 8.1.0
(bundled as a default for a new version of Composer).
### Are you willing to submit PR?
- [X] Yes I am willing to submit a PR!
### Code of Conduct
- [X] I agree to follow this project's [Code of
Conduct](https://github.com/apache/airflow/blob/main/CODE_OF_CONDUCT.md)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
