Re: Location of test results

2016-10-01 Thread Yingyi Bu
+1. Best, Yingyi On Sat, Oct 1, 2016 at 6:21 PM, Till Westmann wrote: > Hi, > > in a recent change [1] the test result location changed from "a directory > in > the module folder" to "a directory in the target folder". So now > a) the test results will get removed with every

Location of test results

2016-10-01 Thread Till Westmann
Hi, in a recent change [1] the test result location changed from "a directory in the module folder" to "a directory in the target folder". So now a) the test results will get removed with every "mvn clean" and b) we don’t need to exclude the test result directories in .gitignore anymore.

Re: Does Projection affect count() performance?

2016-10-01 Thread Khurram Faraaz
This ASTERIXDB-159 maybe a good place to start with, this was recently fixed. On Sat, Oct 1, 2016 at 10:13 PM, Yingyi Bu wrote: > Wail, > > Can you attach the query plan for query 1? > I tried >count(

Re: Does Projection affect count() performance?

2016-10-01 Thread Yingyi Bu
PS, if you still have the OOM instance, can you do a Yourkit memory profile? Thanks! Best, Yingyi On Sat, Oct 1, 2016 at 9:43 AM, Yingyi Bu wrote: > Wail, > > Can you attach the query plan for query 1? > I tried >count( for $x in dataset beers >

Re: Does Projection affect count() performance?

2016-10-01 Thread Yingyi Bu
Wail, Can you attach the query plan for query 1? I tried count( for $x in dataset beers return $x ) and got the following plan, which seems OK: -- DISTRIBUTE_RESULT |UNPARTITIONED| exchange -- ONE_TO_ONE_EXCHANGE |UNPARTITIONED| aggregate [$$5]

Re: Does Projection affect count() performance?

2016-10-01 Thread Khurram Faraaz
Wail please file a JIRA to track this issue. On Oct 1, 2016 2:31 PM, "Wail Alkowaileet" wrote: > Hi, > > I know that early projections will enhance the performance. > I just noticed something: > > 1- returning the whole tuple > count( for $x in dataset Tweets > return $x > )

Does Projection affect count() performance?

2016-10-01 Thread Wail Alkowaileet
Hi, I know that early projections will enhance the performance. I just noticed something: 1- returning the whole tuple count( for $x in dataset Tweets return $x ) => Throws an exception Java heap exceeded. (The heap-size is less than the sum of AsterixDB configured memory ... so it's not a