Counting errors
On May 17, 1:01*pm, kittronald wrote:
I try to stay away from array entered formulas since they
collectively slow my computer down.
I have several reasons for avoiding array formulas, but performance is
not one of them.
Comparing SUMPRODUCT(--ISERROR(range)) to the array formula SUM(--
ISERROR(range)), I find that the SUM formula is about 55 times faster
than SUMPRODUCT for a range of 100 cells and about 390 times faster
for a range of 10000 cells on my system[*].
Nevertheless, we are talking about very small times per formula (less
than 2 msec on my computer).
My primary reason for avoiding single-cell array formulas is that they
are error-prone. Often, they will appear to work (return a value
instead of an error) if we press Enter instead of ctrl+shift+Enter,
resulting in a non-array formula.
-----[*] XL2003 on WinXP. Single-core 2.127 GHz Pentium M processor.
|