Marginally less overhead if the data is sorted:

 perl -lane 'print unless $F[0] eq $prev;$prev=$F[0]' < file2.txt > fi
+le1.txt
[download]

Also, the OP's code handles the boundary condition for the first record, but not the last.