I need to map a large flat-file (over 20MB) into XML.
I am using the ffIterator feature in pub.flatFile:convertoToValues to optimize memory usage.
The input is read using pub.file:getFile service.
The question …
Does it matter whether I read the input as bytes or stream? Also, I see that I need to leave the “body” document (output of pub.file:getFile service) in the pipeline until the last iteration of the REPEAT loop. If I read in as bytes, this would mean I am carrying the whole file in memory throughout? Not sure how it works in case of a read as stream? Could someone please clarify this?
I am thinking that read as stream would be the solution here. But some guidance would certainly help.
Thanks in advance.