In general you’ll need to read with the same transcoder that you wrote with, otherwise you’ll likely encounter an error. So, if you were using the LegacyTranscoder and you switched to the DefaultTranscoder, you may encounter an error as the document was written with different flags.
Yes, possibly, you can always pull down that commit and test it out or wait until we release it the 2nd week of January.
Yeah, for sure.
I do suspect that since you’ve switched transcoders your documents are in a “hybrid” state - or at least some of them are. There are a couple of ways to remedy this:
Flush the bucket and start from scratch with fresh data
Go through the data and read with the converter that works and then write using the transcoder you wish to convert the documents to.
You could even write a custom transcoder and handle the conversion logic there I imagine.
It would be useful to post the entire exception + stacktrace to rule out another issue.