When the source data range contains trailing rows, simply skip them, and
don't even bother iterating them.  Apparently sometimes users specify a
data range with a huge amount of trailing empty rows, which would slow
down the pivot table refresh for no good reason.

But we do still need to keep the original end row position, in case the
pivot table needs to include empty cells in the output.

Change-Id: I2c73c368837b8e322e12b25ddf31429488961f06