Yes, the blob file is not large and thanks again for the help.     ... View more

David,   Thanks a lot for your information. Really appreciate it.   I will see if our team want to invest the time/resources to create a UDF in Hive. Normally, how long it takes to create a UDF like my case? ... View more

In each entry of the CSV file, the last column is the path (HDFS directory path with blob file name) of the blob file. So one blob file is related to each row entry in the CSV file. For the first concern, I just cannot load the blob files into the same Hive table based on the HDFS path. ... View more

Hi David,   Thanks a lot for the response.   The Blob table has no row numbers since we load the Blob files into Hive table from a HDFS directory. So I cannot use JOIN to combine the two tables.   And yes, there is a 1:1 relationship with the number of Blobs to the number of CSV entries.   Your solution is to use JOIN to combine two tables with same rowid but there is a problem that how can we have the rowid for the Blob tables (the row Blob data files are in one HDFS directory). ... View more