2.5.4.12 SCALE Command
  The SCALE command overcomes accumulated round-off
  errors by multiplying double-precision, floating-point values
  by 10 raised to the specified scale factor before storing them.