I think we don't have any concept of indexes in ab inition it is there only in Database like table indexing to improve query performance of the tables.
In Ab initio it is only massive parallel processing concept.
There is as such no concept of indexing in Ab Initio since it is a code based ETL tool. Indexes come into picture on the database table side. But there is something called output indexing in Ab Initio which is carried out using a reformat. Suppose we wish to divide the incoming flow into 5 flows depending on a particular conditions or conditions we increase the number of output flows in the reformat to 5 and then do the necessary coding in transform part. The output gets divided accordingly and this is achieved without having to use multiple reformats.
