https://pinot.apache.org/ logo
Join Slack
Powered by
# troubleshooting
  • d

    Damiano

    05/04/2020, 11:44 PM
    wait
  • d

    Damiano

    05/04/2020, 11:44 PM
    do you mean that i need to execute that return and look at console
  • d

    Damiano

    05/04/2020, 11:44 PM
    right ?
  • s

    Sidd

    05/04/2020, 11:45 PM
    No. When you step through the code in IntermediateResultsBlock in function getDataTable
  • s

    Sidd

    05/04/2020, 11:46 PM
    at that time, see the list of exceptions
  • s

    Sidd

    05/04/2020, 11:47 PM
    how are you debugging this? intellij, gdb via command line?
  • d

    Damiano

    05/04/2020, 11:48 PM
    intellij
  • d

    Damiano

    05/05/2020, 12:44 AM
    The problem has been solved by @Sidd it is basically a timeout error because of the debugging mode... Increasing the timeout value fixed the problem
    👍 1
  • d

    Damiano

    05/05/2020, 7:43 AM
    Hi @Subbu Subramaniam really thank you for asking. At the moment i am debugging my aggregator I will share my code with you very soon. Thanks!
  • d

    Damiano

    05/05/2020, 9:09 AM
    I found a problem with aggregateGroupBySV. It is not accepting more than one column. If i do myAggr(col1, col2) i only get the col1 so it is a similar problem we had in aggregate()
  • d

    Damiano

    05/05/2020, 9:11 AM
    I think something related with AggregationFunctionContext, in the constructor i see columnName = col1:col2 sounds strange, should it works that way?
  • m

    Mayank

    05/05/2020, 1:03 PM
    @Damiano `
    Copy code
    void aggregateGroupBySV(int length, int[] groupKeyArray, GroupByResultHolder groupByResultHolder,
        Map<String, BlockValSet> blockValSets);
  • m

    Mayank

    05/05/2020, 1:04 PM
    This method gets multiple columns' blockvalset via the last arg
  • m

    Mayank

    05/05/2020, 1:05 PM
    In AggregationFunctionFactory, you have a list of args that you can pass to the constructor of you aggr func
  • d

    Damiano

    05/05/2020, 1:09 PM
    yeah i know that i should get all the columsn there...but i am getting only one, the first of myAggr(col1, col2)... regarding the: AggregationFunctionFactory i already changed it in:
  • d

    Damiano

    05/05/2020, 1:09 PM
    Copy code
    case DRAWDOWN:
      if (arguments.size() != 2)
        throw new IllegalArgumentException();
      return new DrawdownAggregationFunction(arguments.get(0), arguments.get(1));
  • d

    Damiano

    05/05/2020, 1:10 PM
    i am getting the columns in the constructor, the problem is when Pinot calls aggregateGroupBySV
  • d

    Damiano

    05/05/2020, 1:11 PM
    Copy code
    BlockValSet blockValSet = blockValSetMap.get(_column);
    blockValSet is null when i try to read the content of the second column, understand?
  • d

    Damiano

    05/05/2020, 1:12 PM
    @Mayank ^
  • m

    Mayank

    05/05/2020, 1:20 PM
    Sorry some of these changes are in the PR: https://github.com/apache/incubator-pinot/pull/5316
  • m

    Mayank

    05/05/2020, 1:20 PM
    Can you refer to how I handle multiple args in
    DistinctCountThetaSketch
    in the PR above?
  • d

    Damiano

    05/05/2020, 1:24 PM
    a lot to change
  • d

    Damiano

    05/05/2020, 1:25 PM
    what can i do @Mayank do i have to change all of those things?
  • d

    Damiano

    05/05/2020, 1:25 PM
    or merge your PR and then add my code again
  • m

    Mayank

    05/05/2020, 1:27 PM
    You can either merge that PR and rebase you’re changes on top. Or You just look inside of the function I gave you to see how it gets multiple columns, and get the idea from therr
  • m

    Mayank

    05/05/2020, 1:28 PM
    If you rebase, I recommend doing it in a new branch so that you keep your original changes intact in the original branch
  • d

    Damiano

    05/05/2020, 1:30 PM
    hmm ok, however i did not create a custom Executor, i just touched DefaultAggregationExecutor.java and groupby/DefaultGroupByExecutor.java
  • d

    Damiano

    05/05/2020, 1:30 PM
    do no need a custom Executor for my aggragator?
  • m

    Mayank

    05/05/2020, 1:30 PM
    You don’t need
  • d

    Damiano

    05/05/2020, 1:40 PM
    @Mayank so i only need to change those files to be able to read multiple columns
1...969798...166Latest