Parsing a vertically strctured flat file

I want to parse a “vertically structured” file so if I have the following - A PO wraps a header (PH) and line item (PL) Is it possible to have the fields of the header on separate lines based on their position after the PH line without a descriptor? Or do they have to go in fixed width format across the line?

(fields are below record in order they appear without record descriptor):
PO
PH
00002
My Header
PL
1234

Second option (fields are beside record descriptor in fixed format):

PO
PH00002My Header
PL1234

or (every line has a record descriptor):
PO
PH
LN00002
MHMy Header
PL
PLN1234