Closed iamnoah closed 7 years ago
+1
I realize this is a large PR, but can we get some sort of feedback? Is this project still going? Should we start our own fork for something like this?
Yeah I would go with a fork, we don't have bandwidth to deal with this currently and are not actually using this internally anymore.
We have a topic that contains Thrift serialize messages, so we wrote a binary format. Since it can be hard to grok binary files, this PR also includes some basic helper classes for reading the files back out of S3.
This also adds support for arbitrary output formats by implementing the Connect Converter interface. The binary format is just one implementation, as is the default newline delimited format.