Github user manishgupta88 commented on the issue: https://github.com/apache/carbondata/pull/2191 @ravipesala During compaction we retrieve dictionary values in the form of MDKey and no dictionary columns as byte array. I can make use of CarbonColumnVectorImpl but in that case I will have to write a new iterator which will fill the data column wise and then combine all dictionary MDKeys into one array and no dictionary data into one array which is input to CarbonRow. This will again have cost of combining the values. Is that ok?
---