Note that you don't need to actually post much data. Strip it down to show the salient variations. Any text that isn't important for identifying records or data items or parameters can be omitted or abbreviated as appropriate. Background like how the original file is generated may be helpful to give us some idea of how much variation there may be in elements of the format. The smaller you can make the data and still be clear about the actual file format the easier it will be to figure out how to parse it.

If the sample we have already is representative of the actual data (exact naming format aside) we don't need large lumps of additional data. It is important that anything related to record breaks is accurately represented though.

Updates to posts which had already so many replies are confusing.

Explain why and use readmore and code

here some more accurate input data ...

<readmore>
<code>
DATA
</code>
</readmore>
[download]

update

°) i.e. your root node