Github user jackylk commented on a diff in the pull request:
https://github.com/apache/carbondata/pull/2374#discussion_r196509935
--- Diff:
hadoop/src/main/java/org/apache/carbondata/hadoop/api/CarbonFileInputFormat.java
---
@@ -174,9 +174,15 @@ public CarbonTable
getOrCreateCarbonTable(Configuration configuration) throws IO
List<InputSplit> result = new LinkedList<InputSplit>();
// for each segment fetch blocks matching filter in Driver BTree
- List<CarbonInputSplit> dataBlocksOfSegment =
- getDataBlocksOfSegment(job, carbonTable, filterResolver,
matchedPartitions,
- validSegments, partitionInfo, oldPartitionIdList);
+ List<CarbonInputSplit> dataBlocksOfSegment;
+ if (carbonTable.getTableInfo().getFormat().equals("")
--- End diff --
why support empty string?
---