walterddr commented on code in PR #9597:
URL: https://github.com/apache/pinot/pull/9597#discussion_r997602810
##########
pinot-query-runtime/src/main/java/org/apache/pinot/query/runtime/operator/operands/FilterOperand.java:
##########
@@ -95,6 +114,36 @@ public Boolean apply(Object[] row) {
@Override
public abstract Boolean apply(Object[] row);
+ private static class BooleanInputRef extends FilterOperand {
+ private final RexExpression.InputRef _inputRef;
+
+ public BooleanInputRef(RexExpression.InputRef inputRef, DataSchema
dataSchema) {
+
Preconditions.checkState(dataSchema.getColumnDataType(inputRef.getIndex())
+ == DataSchema.ColumnDataType.BOOLEAN);
+ _inputRef = inputRef;
+ }
+
+ @Override
+ public Boolean apply(Object[] row) {
+ return row[_inputRef.getIndex()] instanceof Integer ? (Integer)
row[_inputRef.getIndex()] > 0
Review Comment:
yeah I wasn't able to find a way to avoid this in this PR. but I have
another issue open
basically the idea is:
- currently datatable returns the row as stored data type and it doesn't
convert it back to the original data type. for example for data that are
timestamp and boolean, we convert them in the result table step.
- this should not be the behavior for data block. as the read out to
object[] step reside in datablockutils. it should return the proper
columndatatype instead.
once this issue is addressed we no longer need this conversion (in fact we
no longer need conversion of any kind)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
