Open austinweisgrau opened 1 week ago
A bug was introduced by recent PR #1062
If our source table is loaded from CSV with no transformations, the original source file will be directly loaded to GCS.
We may need to pass along a custom delimiter to BigQuery.
e.g. for this workflow:
tbl = Table.from_csv(filepath, delimiter='\t') BigQuery().copy(tbl, 'my_dataset.my_table')
The copy job will fail because the delimiter for the file loaded to GCS will be a \t, but BigQuery().copy() expects a comma by default.
\t
This change fixes that to pass along any custom delimiter set on the Table file load to the BigQuery copy job config.
A bug was introduced by recent PR #1062
If our source table is loaded from CSV with no transformations, the original source file will be directly loaded to GCS.
We may need to pass along a custom delimiter to BigQuery.
e.g. for this workflow:
The copy job will fail because the delimiter for the file loaded to GCS will be a
\t
, but BigQuery().copy() expects a comma by default.This change fixes that to pass along any custom delimiter set on the Table file load to the BigQuery copy job config.