Problems with large fixed-width files

Question

I am trying to read a flat file with a fixed width using JRecordBind. Each line can be a sub-correction of the previous one, and the first line and the last are the header and footer of the root record.

Example:

xxxxxxxxxxxxx -> Record1 header with list of Records2
yyyyyyyyyyy1h -> Record2 instance 1header with list of Records3
@@@@@@@@@@@@1 -> Record3 instance1
@@@@@@@@@@@@2 -> Record3 instance2
yyyyyyyyyyy1f -> Record2 footer
yyyyyyyyyyy2h
@@@@@@@@@@@@3
@@@@@@@@@@@@4
yyyyyyyyyyy2f
zzzzzzz   

My problem is the file size with many records (in my case 140,000), a StackOverFlow exception is thrown.

I am analyzing BeanIO as an alternative to JRecordBind.

What is the best option with these files?