[Mondrian] LastNonEmpty CellRequestQuantumExceededException
Hilario Fernandes
hilario.fernandes at cortex-intelligence.com
Thu Nov 27 06:12:17 EST 2014
Hi Matt,
Thank you for your answer!
About the Aggregate, you are right! It is not needed indeed, it was part of
a bigger MDX and i didn't remove it as i striped it down to a simpler
version.
Unfortunately i do need to calculate the last data month per city in this
case, so the set approach is not suitable. As you said, the exception is
not really an exception, it is a control flag of some sort. The problem
apparently is that it's not being properly handled on the UDF environment.
We managed to turn the UDF into a Fundef with pretty much the same
algorithm and the problem seems to be gone, i believe the flag is properly
handled there.
Thanks
On Wed, Nov 26, 2014 at 2:41 PM, Matt Campbell <mcampbell at pentaho.com>
wrote:
>
>
> Hi Hilario,
>
>
>
> CellRequestQuantumExceededException gets thrown when the number of cell
> requests exceeds the defined threshold (100K by default). It’s not really
> an exception, though. It’s a way of short-circuiting evaluation so that the
> already batched set of requests can be processed.
>
>
>
> Given that your query is pulling 200 years of month level members within
> the LNE, and then evaluating that in the context of each of a few hundred
> cities, I guess it’s not surprising that you would exceed the threshold.
>
>
>
> Just as an aside, there’s couple of things I note with your query: first,
> you’re doing an Aggregate() of what LastNonEmpty returns. LNE returns a
> single member, so there’s no reason to Aggregate. You could just use a
> tuple. Second, the query is grabbing the LNE month in the context of each
> city. This incurs the performance overhead of computing the LNE for each
> specific city (which may be exactly what you want and need). If you really
> only need the LNE month overall, and then to see the Cities price within
> that month, then you could put the LNE within a named set. That will cause
> the last month to be evaluated once in the context of the slicer, and will
> avoid much of the performance overhead.
>
>
>
> HTH
>
>
>
>
>
>
>
> *From:* mondrian-bounces at pentaho.org [mailto:mondrian-bounces at pentaho.org]
> *On Behalf Of *Hilario Fernandes
> *Sent:* Wednesday, November 19, 2014 1:17 PM
> *To:* Mondrian mailing list
> *Subject:* [Mondrian] LastNonEmpty CellRequestQuantumExceededException
>
>
>
> Hi everyone,
>
>
>
> I'm getting the CellRequestQuantumExceededException thrown as an error,
> and sometimes as a warn dependind on the MDX query. This happens while
> using LastNonEmpty and specifically when the set given as argument is
> larger.
>
>
>
> For example, the following query:
>
>
>
> WITH
>
> MEMBER [Measures].[LastMonthPrice] AS
>
> Aggregate(LastNonEmpty([Data.YEAR_MONTH].Members, [Measures].[Price_SUM]),
> [Measures].[Price_SUM])
>
> SELECT
>
> NON EMPTY {[Measures].[LastMonthPrice]} ON COLUMNS,
>
> NON EMPTY {[City].[City].Members} ON ROWS
>
> FROM [Cube]
>
>
>
> I get several warnings like this:
>
>
>
>
>
> WARN olap.ResultBase Mondrian: exception in executeStripe.
>
> mondrian.olap.fun.MondrianEvaluationException: Expected value of type
> MEMBER; got value 'mondrian.olap.fun.MondrianEvaluationException: Exception
> while executing function LastNonEmpty:
> mondrian.rolap.agg.CellRequestQuantumExceededException
>
> at
> mondrian.rolap.agg.CellRequestQuantumExceededException.<clinit>(CellRequestQuantumExceededException.java:37)
>
> at
> mondrian.rolap.FastBatchingCellReader.recordCellRequest(FastBatchingCellReader.java:177)
>
> at
> mondrian.rolap.FastBatchingCellReader.get(FastBatchingCellReader.java:154)
>
> at
> mondrian.rolap.RolapEvaluator.evaluateCurrent(RolapEvaluator.java:661)
>
> at
> mondrian.calc.impl.MemberValueCalc.evaluate(MemberValueCalc.java:100)
>
> at
> mondrian.olap.fun.UdfResolver$CalcExp.evaluateScalar(UdfResolver.java:311)
>
> at
> mondrian.udf.LastNonEmptyUdf.execute(LastNonEmptyUdf.java:70)
>
> at
> mondrian.olap.fun.UdfResolver$ScalarCalcImpl.evaluate(UdfResolver.java:198)
>
> at
> mondrian.calc.impl.GenericCalc.evaluateMember(GenericCalc.java:149)
>
> at
> mondrian.olap.fun.SetFunDef$SetListCalc$2.evaluateVoid(SetFunDef.java:166)
>
> at
> mondrian.olap.fun.SetFunDef$SetListCalc.evaluateList(SetFunDef.java:194)
>
> at
> mondrian.calc.impl.AbstractListCalc.evaluate(AbstractListCalc.java:65)
>
> at
> mondrian.olap.ExpCacheDescriptor.evaluate(ExpCacheDescriptor.java:100)
>
> at
> mondrian.rolap.RolapEvaluator.getCachedResult(RolapEvaluator.java:955)
>
> at
> mondrian.olap.fun.CacheFunDef$1.evaluate(CacheFunDef.java:63)
>
> at
> mondrian.calc.impl.GenericIterCalc.evaluateList(GenericIterCalc.java:55)
>
> at
> mondrian.olap.fun.AbstractAggregateFunDef.evaluateCurrentList(AbstractAggregateFunDef.java:77)
>
> at
> mondrian.olap.fun.AggregateFunDef$AggregateCalc.evaluate(AggregateFunDef.java:102)
>
> at
> mondrian.rolap.RolapEvaluator.evaluateCurrent(RolapEvaluator.java:688)
>
> at
> mondrian.calc.impl.MemberArrayValueCalc.evaluate(MemberArrayValueCalc.java:77)
>
> at mondrian.olap.fun.FunUtil.evaluateMembers(FunUtil.java:418)
>
> at
> mondrian.olap.fun.FunUtil.partiallySortMembers(FunUtil.java:715)
>
> at
> mondrian.olap.fun.TopBottomCountFunDef$3.partiallySortList(TopBottomCountFunDef.java:145)
>
> at
> mondrian.olap.fun.TopBottomCountFunDef$3.evaluateList(TopBottomCountFunDef.java:109)
>
> at
> mondrian.calc.impl.AbstractListCalc.evaluateIterable(AbstractListCalc.java:71)
>
> at mondrian.rolap.RolapResult.executeAxis(RolapResult.java:880)
>
> at mondrian.rolap.RolapResult.evalLoad(RolapResult.java:705)
>
> at mondrian.rolap.RolapResult.loadMembers(RolapResult.java:661)
>
> at mondrian.rolap.RolapResult.<init>(RolapResult.java:289)
>
> at
> mondrian.rolap.RolapConnection.executeInternal(RolapConnection.java:672)
>
> at
> mondrian.rolap.RolapConnection.access$000(RolapConnection.java:52)
>
> at
> mondrian.rolap.RolapConnection$1.call(RolapConnection.java:623)
>
> at
> mondrian.rolap.RolapConnection$1.call(RolapConnection.java:622)
>
> at java.util.concurrent.FutureTask.run(FutureTask.java:262)
>
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
>
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
>
> at java.lang.Thread.run(Thread.java:744)
>
> ' (class mondrian.olap.fun.MondrianEvaluationException)
>
> at mondrian.olap.fun.FunUtil.newEvalException(FunUtil.java:82)
>
> at
> mondrian.rolap.RolapEvaluator.newEvalException(RolapEvaluator.java:1003)
>
> at
> mondrian.calc.impl.GenericCalc.evaluateMember(GenericCalc.java:153)
>
> at
> mondrian.olap.fun.SetFunDef$SetListCalc$2.evaluateVoid(SetFunDef.java:166)
>
> at
> mondrian.olap.fun.SetFunDef$SetListCalc.evaluateList(SetFunDef.java:194)
>
> at
> mondrian.calc.impl.AbstractListCalc.evaluate(AbstractListCalc.java:65)
>
> at
> mondrian.olap.ExpCacheDescriptor.evaluate(ExpCacheDescriptor.java:100)
>
> at
> mondrian.rolap.RolapEvaluator.getCachedResult(RolapEvaluator.java:955)
>
> at
> mondrian.olap.fun.CacheFunDef$1.evaluate(CacheFunDef.java:63)
>
> at
> mondrian.calc.impl.GenericIterCalc.evaluateList(GenericIterCalc.java:55)
>
> at
> mondrian.olap.fun.AbstractAggregateFunDef.evaluateCurrentList(AbstractAggregateFunDef.java:77)
>
> at
> mondrian.olap.fun.AggregateFunDef$AggregateCalc.evaluate(AggregateFunDef.java:102)
>
> at
> mondrian.rolap.RolapEvaluator.evaluateCurrent(RolapEvaluator.java:688)
>
> at
> mondrian.rolap.RolapResult.executeStripe(RolapResult.java:1077)
>
> at
> mondrian.rolap.RolapResult.executeStripe(RolapResult.java:1227)
>
> at
> mondrian.rolap.RolapResult.executeStripe(RolapResult.java:1227)
>
> at mondrian.rolap.RolapResult.executeBody(RolapResult.java:917)
>
> at mondrian.rolap.RolapResult.<init>(RolapResult.java:474)
>
> at
> mondrian.rolap.RolapConnection.executeInternal(RolapConnection.java:672)
>
> at
> mondrian.rolap.RolapConnection.access$000(RolapConnection.java:52)
>
> at
> mondrian.rolap.RolapConnection$1.call(RolapConnection.java:623)
>
> at
> mondrian.rolap.RolapConnection$1.call(RolapConnection.java:622)
>
> at java.util.concurrent.FutureTask.run(FutureTask.java:262)
>
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
>
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
>
> at java.lang.Thread.run(Thread.java:744)
>
>
>
>
>
>
>
> City has a few hundred members and the date is considering dates between
> 1900 and 2100.
>
>
>
> Any ideas?
>
>
>
> Thanks
>
>
>
>
>
> --
>
> Hilario Fernandes
>
> _______________________________________________
> Mondrian mailing list
> Mondrian at pentaho.org
> http://lists.pentaho.org/mailman/listinfo/mondrian
>
>
--
Hilario Fernandes
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.pentaho.org/pipermail/mondrian/attachments/20141127/29f1abcd/attachment-0001.html
More information about the Mondrian
mailing list