In my project, I have to design a Job which should read a file, validate and load into DB2 tables. There are 2 different file volumes i.e. one comes with 5 - 10 million records (of LRECL 500 approx.) and other file with 20K to 100,000 records (of LRECL 400 approx.). I have to decide on either having single program which does all validation and load data directly into Table thro Multi-row Insert, Or having a program which prepares load dataset and then thro BMCLOAD load utility inserting them into the Table. Can you please suggest me which approach will be best (if possible with data facts)?