- Added a new append_only config to allow for only inserting new rows and not updating existing data
- Fix: Instead of failing, use only the first 4 primary keys when clustering if more than 5 primary keys are defined by the tap stream's schema.
- Refactoring code to make it easier to add improvements
- Support for microseconds
- Fixes for date fields
- Support for ACTIVATE_VERSION
- Fixes for clustering when to Primary Key present
- Fix for taps not sending the
time_extracted
field
- Add
hard_delete_mapping
for switching specific streams - Add
batch_wait_limit_seconds
config - Fix nested keys issue
- bump multiple dependencies
- Allow
REPEATED
fields of typeRECORD
to be created. - bump pylint
- bump joblib to 1.1.0 to fix issue #7
- bump test dependencies
- remove unused code
- Minor internal bugfixes
- Add location BigQuery configuration
- Fix for reserved table names
- Fix for bad decimals
- Improved logging of bad data
- Refactor code to make it more similar to other taps
- Truncate big numbers to the maximum range for NUMERIC in BigQuery
- Fix an issue when loading table names with reserved keywords
- Fixed an issue when the source tap can include uppercase primary keys
- Initial release