Sounds like the file was converted to a DOS file when moved between environments.

My first suggestion, normalize the table.  Right now, if you add another step you have to add another column.  If you normalize the data each step becomes a row in...

That is a Unix file terminated only be <lf> character.  Have you tried only using that as the row delimiter?

Looks like Luis and I were at least thinking along the same lines, just a slightly different solution.

I came up with this:

SELECT
  [p].[Product]
  , [p].[Type1]
  , [p].[Type2]
  , [p].[Type3]
  , [p].[Type4]
  , [lt].[Product]
  , [lt].[Type1]
 ...

Out of curiosity, have you looked at the files using a hex editor?

I found this: https://blogs.technet.microsoft.com/rmilne/2012/07/19/using-a-pal-template-to-easily-capture-performance-data/

I have worked with PAL a few years ago but not with a template.  Have thought about looking at it again but haven't had...