Indhumathi27 commented on a change in pull request #3778:
URL: https://github.com/apache/carbondata/pull/3778#discussion_r480961113



##########
File path: 
integration/spark/src/main/java/org/apache/spark/sql/secondaryindex/query/SecondaryIndexQueryResultProcessor.java
##########
@@ -243,46 +258,115 @@ private void 
processResult(List<CarbonIterator<RowBatch>> detailQueryResultItera
   /**
    * This method will prepare the data from raw object that will take part in 
sorting
    */
-  private Object[] prepareRowObjectForSorting(Object[] row) {
+  private Object[] prepareRowObjectForSorting(Object[] row,
+      Map<Integer, GenericQueryType> complexDimensionInfoMap, int[] 
complexColumnParentBlockIndexes)
+      throws SecondaryIndexException {
     ByteArrayWrapper wrapper = (ByteArrayWrapper) row[0];
-    // ByteBuffer[] noDictionaryBuffer = new ByteBuffer[noDictionaryCount];
+    byte[] implicitColumnByteArray = wrapper.getImplicitColumnByteArray();
 
     List<CarbonDimension> dimensions = segmentProperties.getDimensions();
     Object[] preparedRow = new Object[dimensions.size() + measureCount];
+    Map<Integer, Object[]> complexDataMap = new HashMap<>();
 
     int noDictionaryIndex = 0;
     int dictionaryIndex = 0;
+    int complexIndex = 0;
     int i = 0;
     // loop excluding last dimension as last one is implicit column.
     for (; i < dimensions.size() - 1; i++) {
       CarbonDimension dims = dimensions.get(i);
-      if (dims.hasEncoding(Encoding.DICTIONARY)) {
+      boolean isComplexColumn = false;
+      // check if dimension is a complex data type
+      if (!complexDimensionInfoMap.isEmpty() && 
complexColumnParentBlockIndexes.length > 0) {

Review comment:
       dim is actually the index table dimension, where complex type of main 
table is stored as its primitive type in SI. Hence, we cannot check isComplex 
from dimension.




----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


Reply via email to