Jesse, that's very valuable. Let me explain the reason of my interest. I am writing a book "HBase Design Patterns" for Packt, and I am using Phoenix as the way to explain them, rather than making people read through Java code. I go as far as to generate SQL statements with Java code and then run them with Phoenix. I would like to be as precise and as complete about Phoenix as I can.
Thank you. Sincerely, Mark On Tue, Feb 25, 2014 at 10:02 PM, Jesse Yates <[email protected]>wrote: > I worked on adding tracing into phoenix, but never finished it up*; it > would use Cloudera'a HTrace library since that is already bundled into > HBase 0.96+. With the 4.0 phoenix branch (which is based on HBase 0.96+) > this integration should be even easier, and might actually get finished! > > Until then, explain is as far you can take it. > > *https://github.com/jyates/phoenix/tree/tracing It worked at the time on > Hadoop2, but never quite brought it the extra mile necessary to roll it > into phoenix proper. Interesting aside, its also been the main reason I > made the build multi-module (and included the hadoop-compat stuff). > > ------------------- > Jesse Yates > @jesse_yates > jyates.github.com > > > On Tue, Feb 25, 2014 at 7:28 PM, James Taylor <[email protected]>wrote: > >> Yes, see http://phoenix.incubator.apache.org/language/index.html#/explain >> >> It's not a trace, though, it's an execution plan. We don't have a >> profiler yet, though there's been a bit of work on that through the metrics >> framework that HBase provides. >> >> Thanks, >> James >> >> >> On Tue, Feb 25, 2014 at 7:11 PM, Mark Kerzner <[email protected]>wrote: >> >>> Hi, >>> >>> In Cassandra, there is an option >>> >>> trace on >>> >>> which allows you to profile the execution of your CQL statements. Is >>> there a similar option in Phoenix? >>> >>> Thank you for that, and for all the answers before. >>> >>> Sincerely, >>> Mark >>> >>> >> >
