Changelog

1.2.0

Implement multi-processing to improve extraction performance #20

1.1.4

Handle non-union Avro types correctly during sync #18

1.1.3

Only send activate version if we sent records
Only filter files on the bookmark if the bookmark is in the files list
Sort the files explicitly using the dump id and part number instead of using string sorting
#16

1.1.2

Handle case where there are no incremental=false dumps #14

1.1.1

Check for a version bookmark and if it is not present then do a full table resync #12

1.1.0

Updated how we read manifest files and how we choose what files to sync. Added a version bookmark and tests. #11

1.0.2

Add logging of the dump id / manifest id, files, and incremental value when full sync is requested based on heap's documentation of incremental #9

1.0.1

Add logging around Activate Version Messages

1.0.0

Release as open beta

0.0.7

Use filter_data_by_metadata() instead of transform() to speed up record throughput #7