clintropolis commented on code in PR #13803:
URL: https://github.com/apache/druid/pull/13803#discussion_r1146897221
##########
processing/src/main/java/org/apache/druid/segment/NestedDataColumnIndexer.java:
##########
@@ -59,16 +62,43 @@ public class NestedDataColumnIndexer implements
DimensionIndexer<StructuredData,
protected final StructuredDataProcessor indexerProcessor = new
StructuredDataProcessor()
{
@Override
- public ProcessedLiteral<?> processLiteralField(ArrayList<NestedPathPart>
fieldPath, Object fieldValue)
+ public ProcessedLiteral<?> processLiteralField(ArrayList<NestedPathPart>
fieldPath, @Nullable Object fieldValue)
{
- final String fieldName =
NestedPathFinder.toNormalizedJsonPath(fieldPath);
- LiteralFieldIndexer fieldIndexer = fieldIndexers.get(fieldName);
- if (fieldIndexer == null) {
- estimatedFieldKeySize +=
StructuredDataProcessor.estimateStringSize(fieldName);
- fieldIndexer = new LiteralFieldIndexer(globalDictionary);
- fieldIndexers.put(fieldName, fieldIndexer);
+ // null value is always added to the global dictionary as id 0, so we
can ignore them here
+ if (fieldValue != null) {
+ // why not
+ final String fieldName =
NestedPathFinder.toNormalizedJsonPath(fieldPath);
+ ExprEval<?> eval = ExprEval.bestEffortOf(fieldValue);
+ LiteralFieldIndexer fieldIndexer = fieldIndexers.get(fieldName);
+ if (fieldIndexer == null) {
+ estimatedFieldKeySize +=
StructuredDataProcessor.estimateStringSize(fieldName);
+ fieldIndexer = new LiteralFieldIndexer(globalDictionary);
+ fieldIndexers.put(fieldName, fieldIndexer);
+ }
+ return fieldIndexer.processValue(eval);
}
- return fieldIndexer.processValue(fieldValue);
+ return StructuredDataProcessor.ProcessedLiteral.NULL_LITERAL;
+ }
+
+ @Nullable
+ @Override
+ public ProcessedLiteral<?> processArrayOfLiteralsField(
+ ArrayList<NestedPathPart> fieldPath,
+ Object maybeArrayOfLiterals
+ )
+ {
+ final ExprEval<?> maybeLiteralArray =
ExprEval.bestEffortOf(maybeArrayOfLiterals);
+ if (maybeLiteralArray.type().isArray() &&
maybeLiteralArray.type().getElementType().isPrimitive()) {
+ final String fieldName =
NestedPathFinder.toNormalizedJsonPath(fieldPath);
+ LiteralFieldIndexer fieldIndexer = fieldIndexers.get(fieldName);
+ if (fieldIndexer == null) {
+ estimatedFieldKeySize +=
StructuredDataProcessor.estimateStringSize(fieldName);
+ fieldIndexer = new LiteralFieldIndexer(globalDictionary);
+ fieldIndexers.put(fieldName, fieldIndexer);
+ }
+ return fieldIndexer.processValue(maybeLiteralArray);
+ }
+ return null;
Review Comment:
i modified this stuff a bit and updated javadocs so it is hopefully clearer,
the new abstract method names are `processField` and `processArrayField` and
the latter indicates that returning a non-null value halts further processing
of arrays, otherwise the processor will continue for each element of the array
##########
processing/src/main/java/org/apache/druid/segment/NestedDataColumnIndexer.java:
##########
@@ -145,6 +175,10 @@ public DimensionSelector makeDimensionSelector(
final int dimIndex = desc.getIndex();
final ColumnValueSelector<?> rootLiteralSelector =
getRootLiteralValueSelector(currEntry, dimIndex);
if (rootLiteralSelector != null) {
+ final LiteralFieldIndexer root =
fieldIndexers.get(NestedPathFinder.JSON_PATH_ROOT);
+ if (root.getTypes().getSingleType().isArray()) {
+ throw new UnsupportedOperationException("Not supported");
+ }
Review Comment:
clarified exception
--
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]
