Owen O'Malley

The Presto team has some code for HLL. Format description - https://github.com/airlift/airlift/blob/master/stats/docs/hll.md Code - https://github.com/airlift/airlift/tree/master/stats/src/main/java/io/airlift/stats/cardinality I need to play with it, but the summaries can be pretty large.

The problem that I'm having is that I'd like the SparkOrcWriter to get the table properties without looking up the table again. In particular, it would be good to be...

I wasn't clear on what was required, I just see the reader path for Parquet and Avro mangle the schemas for partitioned tables: ```java if (hasJoinedPartitionColumns) { // schema used...

The extra thread is used so that if the calling thread is interrupted while the request is being written to the socket, that the write is not interrupted. Basically, it...

My patch also guarantees that the requests are written in the order they arrived, which the current code does not.

@steveloughran Can you recommend someone to review this RPC patch?

@goiri No, I committed the change to trunk and branch-3.3. Was there still an open issue on this PR?

We need to leave the pure java library as an option with automatic fallback because not every platform will have the native library. Furthermore, if there is a bug in...

Relative paths would help us too. We certainly need to have support for a federated namespace with a virtual file system. We're moving to HDFS federation and moving our tables...

Sorry, I got busy with other work. I've included fixes that we've deployed for a while now. I had missed some cases of IOException handling and start up issues.