I have a flat file csv format that has a header, and line items. The records have a record identifier in the first position H for header and L for line item. The way I process the file now with java is that I read from header record to header record, so I read a header record and then read each line after that record which is reading in the lines until I read another header which then is the start of the next record. So I don't have an explicit end of record marker. I am trying to convert the process over to Spring Batch and I see that I can implement a RecordSeperatorPolicy which will call the isEndOfRecord method to determine if all the lines that make up the record have been read in. However since I have no explicit record terminator I am not sure how what I would need to do in order to implement the RecorSeperatorPolicy to work in my situation.