https://github.com/turbolytics/sql-flow
This allows for ~10x higher throughput, over the python core, and more predictable memory usage! This also makes it easier for our users to run in resource constrained environments (such as IoT).
The go rewrite also brings a bit more sanity to handling background sqlflow control loops, such as window compaction and timeouts.
I'd really love and appreciate your feedback, any examples using sqlflow in production, or your concerns on what makes sqlflow UNUSABLE for you!
Thank you
Danny danny@turbolytics.io