Apache Pinot

No, likely you won't be able to just take one file.

You should have your changes on top of this PR, I think Kishore sent you a link on how to do that in the <#CDRCA57FC|> channel

<@UDQU92KBK> at the moment instead of loading the entire PR i am trying to change the files i mentioned and see what happen

do you know why you are not getting the columns? See how the blockValMap is created in DefaultGroupByExecutor

May be you need to do the same hack you did in DefaultAggregationExecutor

if i will be able to run correctly this indicator i will do tons of other useful aggregators that i will share with you

<@UDQU92KBK> yes but as you can see in GroupBy there is not about DISTINCT, do you remember?

```if (AggregationFunctionUtils.isDistinct(functionContexts) || AggregationFunctionUtils.isDrawdown(functionContexts)) { .. }```


we added that OR with a simple return true in isDrawdown() method

in groupby i did not find anything similar

i lose the secondo columns where:  `_aggregationExpressions[i] = TransformExpressionTree.compileToExpressionTree(functionContexts[i].getColumnName());`

but then in the aggregationExpressions i only see the reference to the first column only

Sorry I cannot think of an easy hack for this one

now iam starting copying your code and see what happen

```    if (AggregationFunctionUtils.isDistinct(functionContexts)) {
      // handle distinct (col1, col2..) function
      // unlike other aggregate functions, distinct can work on multiple columns
      // so we need to build expression tree for each column
      _functions[0] = functionContexts[0].getAggregationFunction();
      _resultHolders[0] = _functions[0].createAggregationResultHolder();

      List&lt;String&gt; expressions = functionContexts[0].getExpressions();
      _expressions = new TransformExpressionTree[expressions.size()];

      for (int i = 0; i &lt; _expressions.length; i++) {
        _expressions[i] = TransformExpressionTree.compileToExpressionTree(expressions.get(i));

      }```

If your query has only one aggregation function, you can pass the columns in _expression[0] and _expression[1] -&gt; the same way as previous case

<@UDQU92KBK> what is StarTreeGroupByExecutor exactly?