Occasionally we deal with data that is so uncooperative that we might as well give up and go back to calculators & ledger books.
Recently I found myself in such a situation and learned something new.
Introducing … data that won’t play nice
Drum roll please. Here is a data-set that I got from somewhere.

The problem – build a lookup formula
And the problem. Oh, simple. Write a lookup formula to find how many customer walk-ins we have on any given day.
But how?
After wrestling with a few variations of VLOOKUP, INDEX, MATCH, OFFSET… I found the right formula for this occasion.
SUMIFS Formula
That’s right, the good old (well, its just 7 years old, born in 2007 Excel) is the one that works in this situation.
Assuming our data is in the range:
- B4:H15
- Even rows (4,6,8…14) containing date
- Odd rows (5,7,9…) containing customer walk-ins
- And lookup date is in L5
We can use this SUMIFS formula:
=SUMIFS(B5:H15,B4:H14,L5)
to find out how many customers walked in to the store on the lookup date.
How does this formula work?
See this illustration.

Since SUMIFS (and SUMIF too) formula works in 2D, our formula works elegantly.
Any caveats?
There are a couple of problems with this formula.
- It works for looking up numbers only
- It doesn’t work if any of the numbers are same as dates we are looking up. For example 15th October 2014 is 41927, so we have a data point that is also 41927, then our SUMIFS result would be wrong.
A better formula?
If you want a formula that works with any type of data, then of course we can come up with one.
That is your home work.
Go ahead and figure it out and share your magic potions in the comments.
Here is a clue: If we cant go somewhere directly, we can go indirectly, if we know the address.
Download Example Workbook
Click here to download the example workbook. Examine the formula in cell L6 to understand how this works.
How do you deal with such data?
Badly shaped data is everywhere. Not a week goes by when I don’t come across poorly structured data. Thanks to Excel, I can alter the shape of almost any data. My favorite techniques are – text import, power query, formulas, pivot tables, remove duplicates & go to special.
What about you? How do you deal with poorly structured data? What tools do you deploy to alter the fitness levels of your data? Please share your thoughts and tips in comments.
Don’t forget about the formula challenge.
Excel not co-operating? Its time you tamed it
Data that is messy, bosses with high expectations, coffee machines that don’t work are everywhere. Fortunately Excel can fix 2 of those 3 problems. (Sorry folks, there is no fixCoffeeMachine() macro, yet.)
Learn a few Excel tricks & concepts and you could be on your way to taming messy data & demanding bosses. Start with these:














17 Responses to “Custom Number Formats – Colors”
You are right, Chandoo. I was playing with the colour numbers last week and some of them don't appear different from each other. Others are totally different from yours.
@Duncan
Each version of Excel, post 2003, renders colors slightly differently
Different language versions may also have different default color palettes
Hello in french
excel 2010
colo1 = couleur1 = black
[couleur1]; [couleur2]; etc..
@Hui, thank you very much again for this great post.
However - under Excel 2007, Hungarian version your solution does not work with color names. I've tried both English and Hungarian names, but drops an error message "not valid formats"
Do you have any idea how to solve this issue?
thanks in advance
@Andras
Without a Hungarian version of Excel 2003 I don't think I can assist
Have you tried using the colour numbers? I couldn't get the names to work (despite using an english version of excel). but it did work with the numbers though. I left out the "u" and was easily able to produce burgundy using [color9]
Here a possible solution: find an English version of Excel, write there the formats using English names, then open the file in the Hungarian version and see the translation.
In Excel 2007 I can't get the colour names to work e.g Sea Green but the numbers do e.g color3 - colour3 does not work so I must bow to the country that has stolen my language (ha ha!)
Hey chandoo, nice Tip!
Wouldn't be easier just apply some conditional formatting for negative numbers and another for positive numbers? Or there's some cases that you can't do that?
Unfortunately the TEXT function doesn't color the cell as number formatting does.
Hi Hui,
Great post Sir, love the new way of formatting with color numbers.
I am using 2007, and it leads me to the last color number 56.
Thanks Hui.
[…] explains how to set up custom number formats with a wide array of […]
Thanks Hui - works a treat!
Thank you, very helpful.
Trying to figure out if it is possible to apply color only to a part of the cell?
E.g. I have a value formatted as Accounting with a currency symbol.
Those I find somewhat distracting though necessary. If I could make them less obtrusive by coloring them gray while the number would stay black, that would be great. Tried tinkering with the format string, but didn't get the desired result. Single color for complete cell value works, but coloring just part of it could not be achieved. Maybe somebody managed that?
Exactly what I was looking for - thank you!
colour in the Australian doesn't work - we have to go American and no problem.
I always thought is was 56 colours notice you have 57. Cool.
thanks
Analir Pisani
Customised Microsoft Office Training Specialist
Sydney - Australia
http://www.azsolutions.com.au
Thank You!