Github user mohammadshahidkhan commented on the issue: https://github.com/apache/carbondata/pull/1819 @geetikagupta16 org/apache/carbondata/spark/rdd/CarbonDataRDDFactory.scala:507 if (loadStatus == SegmentStatus.LOAD_PARTIAL_SUCCESS && status(0)._2._2.failureCauses == FailureCauses.BAD_RECORDS && **carbonLoadModel.getBadRecordsAction.split(",")(1) == LoggerAction.FAIL.name**) The problem is checking equally without ignoring the case. The same problem will arise even with load options option 'bad_records_action'='fail'. Please correct there as well and add test case to cover this scenario.
---