is this mapping question any clearer?

I have a source schema that can have 1-4 PID segments.  However, I need my output to create 4 output records regardless of how many PID’s there are.  How can I simulate more source iterations then there actually are? 

source schema looks like:

PID
  rec1desc
/PID
PID
  rec2desc
/PID

target looks like

rec1
   tag1  (TAGS ARE HARDCODED)
   outputfield1
/rec1
rec2
   tag2  (TAGS ARE HARDCODED)
   outputfield2
/rec2
rec3
   tag3  (TAGS ARE HARDCODED)
   outputfield3
/rec3
rec4
   tag4  (TAGS ARE HARDCODED)
   outputfield4
/rec4

if input data looks like (notice only 2 records here)

pid
   testing1
/pid
pid
   testing2
/pid

then I need output to look like this:(notice the first value in this .csv output was the hardcoded tag value)

Description 1,testing1
Description 2,testing2
Description 3,
Description 4,

the final output file is positional by row so that is exactly why I need this many records generated.  even though there were only two records, the last two just have empty values.