I'm not sure, but this may be safer (and faster) than IF FILE():

CLEAR

LOCAL lcFileToFind, llokay
lcFileToFind = "\mydirectory\mfile.txt"
llokay = findmyfile(lcFileToFind)
? llokay
RETURN

FUNCTION findmyfile(tcFilename)

DECLARE LONG FindFirstFile IN Kernel32 AS funcFindFirst ;
	STRING cFileName, STRING @WIN32_DATA
DECLARE FindClose IN Kernel32 AS funcFindClose LONG hHandle

LOCAL lcFileName, lcBuffer, lcFilePath, llReturn
lcFilePath = ADDBS(JUSTPATH(tcFilename))
lcFileName = JUSTFNAME(tcFilename)
lcBuffer = REPLICATE(CHR(0), 318)
hHandle = funcFindFirst(tcFilename, @lcBuffer)
IF (hHandle > 0)
	lcFileName = SUBSTR(lcBuffer, 45)
	lcFileName = LEFT(lcFileName, AT(CHR(0), lcFileName)-1)
	? lcFileName
	funcFindClose(hHandle)
	llReturn = .T.
ELSE
	? lcFilePath+lcFileName+" Does NOT Exist"
	llReturn = .F.
ENDIF
CLEAR DLLS funcFindFirst, funcFindClose
RETURN llReturn

Also, just as a test, if instead of an APPEND, if you step through the file and copy records one at a time (while slower), does the CPU Utilization still max out? I'm curious because I have seen CPU Utilization jump up to 90% doing an append in the past.