Excel - decimals

  • Thread starter Thread starter school girl
  • Start date Start date
S

school girl

While using Excel, I find that the system isn't rounding appropriately to
whole numbers - i.e., 84.5 becomes 84 - when it should be 85.

Any help would be appreciated!
 
How are you rounding? post the formula.

One possibility you have (say) 8.49 in a cell and are formatted to show 1
decimal place. Excel will display 8.5 but the real value remains as 8.49.

If we round this

=round(a1,0)

Excel worls on the 'real' value of 8.49 and round down to 8
--
Mike

When competing hypotheses are otherwise equal, adopt the hypothesis that
introduces the fewest assumptions while still sufficiently answering the
question.
 
The formula refers to two other cells (e.g., N2+K2/2) which have formula of
their own (e.g., =M26/$M$10) which aer looking at another cell which has a
formula (e.g., =SUM(D26:L26)), and ultimately returns with an answer of 84.5
- which the system rounds to 84 -

I even tried having the system round to the nearest whole number - and it
still rounded to 84 - and unless my elementary school teachers were wrong,
84.5 should round up to 85 not down to 84.

Any additional suggestions?
 
school girl said:
unless my elementary school teachers were wrong,
84.5 should round up to 85 not down to 84.

No need to get snippy. Your school teachers were not wrong. You are wrong
in the way that you are looking at the numbers.

The point that Mike tried to make is: "what you see is NOT what you have".

The formula refers to two other cells (e.g., N2+K2/2) [....]
and ultimately returns with an answer of 84.5
- which the system rounds to 84

It only __appears__ to be 84.5. Format the cell as Number with 13 decimal
places. It will likely show a different number, as large as
84.4999999999999. That is why ROUND(N2+K2/2,0) is not 85.

(If you see 84.5000000000000, there are ways to explain what is going on.
But the explanation is considerably more involved.)

The following work-around should match your expectation:

=ROUND(ROUND(N2+K2/2,1),0)

There might be better things to do in the long-run. But you do not provide
enough details to offer specific guidance.


----- original message -----
 
Back
Top