>I just ran a simple test;
>
>

>LOCAL lnStart, lnEnd
>ldDate = DATE()
>lnStart = SECONDS()
>FOR x = 1 to 100000
>	ldLastDay = GOMONTH( ldDate,1 )
>	ldLastDay = ldLastDay - DAY( ldLastDay )
>ENDFOR
>lnEnd = SECONDS()
>? "2 lines: ", lnEnd - lnStart
>
>lnStart = SECONDS()
>FOR x = 1 to 100000
>	ldLastDay = GOMONTH( ldDate,1 ) - DAY( GOMONTH(ldDate,1))
>ENDFOR
>lnEnd = SECONDS()
>? "1 line: ", lnEnd - lnStart
>

>
>And got the 1 line method to be slightly faster than the 2 line method. The actual times varied on each run but an average of 0.591 seconds for the 2 line version and 0.530 for the one line approach.

That's a very similar approach to the one I used to determine that my EOM() routine was faster than yours.