In our old hadoop cluster, when we create csv file by using spark stored in hdfs temp location and we can cat (read) the csv file. We can see the actual digit/number. But in our new built hadoop cluster, when we pass cat command for csv file, we can not read the file. The file is not human readable. Even though we pass UTF-8 in our code.
Both cluster has same code in spark. Then why one cluster csv file is not readable ?