Excel Tables have been around for a decade now (they are introduced in Excel 2007), and yet, very few people use them. They are versatile, easy and elegant. At Chandoo.org, we celebrate Tables all the time. If you have never used them, start with below tuts.
- Introduction to Excel tables
- How to use structured referencing
- Tables and Relationships in Excel
- Using lookups and other formulas with Excel tables
- Simple way to get absolute references in Tables
- Customizing table styles for awesome usability
While tables are super helpful, they do come with some limitations. Today let’s examine one such unique problem and learn about an elegant solution.
Table Relative Reference Problem
Imagine you are the machine supervisor at Mighty Machine City Inc. Although your machines are mighty, sometimes they do fail. To keep track of which machines are under repair, you maintain a repair log in Excel. Since you heard Tables are mighty, you thought,
‘Gee whiz, I might as well use tables to maintain the repair log. Chandoo says tables are sweet’
So, your Repair log looks like this:
After a few days of tracking the repairs, you wanted to know if same machines are failing successively. For example, in above picture, you notice that MACH-0038 failed twice in a row starting with 11th of March. Same goes for few other rows.
You are the kind of person who frowns upon manually highlighting yellow color in cells to flag such successive failures. So you want to write a formula.
So you add a new column called Same? and want to fill it up with a simple relative reference formula to check with Machine in row 1 matches machine in row 2.
Here is the formula you used:
=[@[Machine ID]] = B6
Note: Your table starts with Row 5.
Excel automatically filled down the formula for all rows of the table, because tables are awesome like that.
You whistled your way to home that night.
Next day morning, as usual Homer messed up something and you had a new repair to log. So you went to the bottom of fail table and inserted a new row to add the failure details.
And you notice something unusual.
The formula for Same? column is WRONG!!!
As soon as you inserted a new row, Excel adjusted last row’s formula to something silly.
- Before: Let’s say the last row formula reads =[@[Machine ID]]=B20
- After: The last but one row (as you now have an empty row)’s formula reads =[@[Machine ID]]=B21
Now that is clearly wrong!
What is going on here?
Your relative referencing worked ok, until the last row. At this stage, Excel understood the formula as Current row value = value in first cell below table
The part in green is what caused trouble. As soon as you add a new row to your table, fist cell below table is moved down. So Excel adjusted that reference alone.
How to fix this problem?
The usual method to fix this:
- Insert as many rows as you need and complete entering / pasting all data.
- Select the formula in very first cell.
- Fill it down all the way (you can double click on the bottom right corner of the cell)
But that is soooo not awesome.
You are right. This method is manual and error prone. It is the opposite of awesome.
Problems when you delete too: In fact, if you ever delete a row from your table, the formulas further down would show #REF! errors. So this method is not very effective in real life.
An elegant way to get relative references in tables
Instead of using cell address based references, like B6, if you use pure structured references, then Excel will automatically adjust them as your table grows or shrinks.
But how?
Simple, we can use OFFSET function along with @ references.
To get next machine ID, you can use
=OFFSET([@[Machine ID]],1,0)
So, to check if same machine failed twice in a row, use
=[@[Machine ID]]=OFFSET([@[Machine ID]],1,0)
As this uses no cell references, whenever you add / change / remove table rows, Excel automatically scales the formula.
But I heard OFFSET is volatile / dangerous / RDX / %#$&@#?
Unless your table has a 200k+ rows or you plan to set up 100s of columns like this, don’t bother. for small, day to day tables, there is no change in performance. If you really hate OFFSET, try talking about it during your next therapy session. Jokes aside, you can also use a longer INDEX based formula to get similar result, but that is semi-volatile too.
Here is one such INDEX based formula.
=INDEX([@[Machine ID]]:INDEX([Machine ID],COUNTA([Machine ID])), 2)
It sure is a mouthful. You can shorten it by using a named formula for INDEX([Machine ID],COUNTA([Machine ID])) portion or the whole thing. Again, I wouldn’t recommend the INDEX based approach over OFFSET for smaller data sets. For larger datasets, see if you can fix the problem at source (for example, modifying your SQL to get offset values in a separate column) or using Power Query to mash the data (more on this in a next post).
So there you go, an elegant and simple way to deal with the relative reference problem in tables.
Bonus tip: generating running numbers in tables
You can use this approach to generate running numbers (1,2,3…) in a table column that grow / change / shrink based on your table. This can be very useful in many scenarios.
To get running numbers in a table column, just use:
=ROWS(fail[[#Headers],[Machine ID]]:[@[Machine ID]])-1
The pattern that you can use in any table goes like this:
=ROWS(Table_Name[[#Headers],[Col 1]]:[@[Col 1]])-1
Bonus Bonus tip: If you have running numbers in a column …
You can then change the OFFSET based relative ref to INDEX like below.
=INDEX([Machine ID],[@Running]+1)
This method is 100% non-volatile, but does return #REF! error for the last row. So use it with IFERROR when nesting in other formulas.
Download Example Workbook
Click here to download an example workbook showcasing all these techniques. Use the extra data to paste and test various methods.
How do you write relative refs in tables?
The handful times when I had to use relative refs in a table, I resorted to cell refs (like B6 above). But this created too much headache further down. So I switched to OFFSET / INDEX approaches.
What about you? How do you write relative references in tables? Please share your stories & struggles in the comments section.
Related Reading (no pun)…
If you are relatively free and want some relaxed reading then check out below related reference links.
- An intro to Excel relative references, structured references
- INDEX formula, OFFSET formula, ROWS formula
Happy learning.
20 Responses to “Untrimmable Spaces – Excel Formula”
Hi Chandoo,
First of all, HAPPY NEW YEAR!!! Wish you and your family another fruitful year ahead.
To answer your question: Power Query is the best way to trim. 🙂
Btw, if Power Query is not available, then formula would absolutely do... but did you forget to mention also Char 32?
One more question: Is the trailing minus meant to be a negative number? Maybe only the sender knows... 🙂
Cheers,
I just see your PQ way, it is amazing, I think it is the most simple way.
No idea how it did it?
I know these spaces can be a real pain but these days I advise Excel users to learn and use Flash Fill and that will learn what to do pretty quickly.
Highlight range to be cleaned. Then, in Replace, hold down the Alt key and type 0160. Replace with nothing.
I accomplished this by writing a macro to go through all the possible unprintable characters. Looped through the range.
@Steve
Brute force works just as well, its just slower
I use a different method here. First, I will copy the data from Excel and paste it in a notepad. In Notepad, I will do a Find Blanks (Space " ") and Replace (Empty) with nothing.
Then you can copy the data from Notepad and paste it back to Excel which will be a perfect number as you desire.
But Thanks for the formula. Its probably the 2nd out of 8 tricks as Chandoo mentioned. Waiting for the rest among 8 from other users 🙂
Hi....
You don't always need notepad for that. I use the Find/Replace is Excel works just fine.
I don't understand the x's. Why weren't they removed in the formula? Or are they part of some sort of numeric formatting that I'm not familiar with? I saw how you handled the non-breaking spaces and the dashes, but am confused about what role the x's played in all this.
Thanks!
Hi Andrew ,
The xs have been used solely to demarcate the actual data text ; thus , without the x in place at the end of text , as in :
x 4,124,500.00 x
it would be impossible to know that there are unwanted trailing characters , in this case , after the last 0.
These xs are not part of the original data text , nor are they used in the formulae ; they are put in only so that readers can visualize the individual items of data as they are in practice. Think of them as imaginary delimiters.
Oh, that makes sense! Thank you for the explanation. I had a feeling it was something along those lines.
You can type this character using the Keys Alt+0160.
Very useful to replace this Character using Find and Select resource.
For many years, my jobs have included ETL tasks and I built this macro to help long, long ago. I tweak it every now and again. Many co-workers, past and present, have it wired to a button on their toolbar.
Sub Clean_and_Trim()
'CAUTION: Strips leading zeroes -- do not use on zipcodes, etc.
If Application.Calculation = xlCalculationAutomatic Then
Application.Calculation = xlCalculationManual
Revert = 1
ElseIf Application.Calculation = xlCalculationManual Then
Revert = 0
End If
For Each Cell In Selection
For x = Len(Cell.Value) To 1 Step -1
If Asc(Mid(Cell.Value, x, 1)) = 160 Then
Cell.Replace What:=Chr(160), Replacement:=" ", LookAt:=xlPart, MatchCase:=True
End If
If Asc(Mid(Cell.Value, x, 1)) = 32 Then
Cell.Replace What:=Chr(32), Replacement:=" ", LookAt:=xlPart, MatchCase:=True
End If
Next x
If Cell.Value "" Then
Cell.Value = Application.Clean(Application.Trim(Cell.Value))
End If
Next
If Revert = 1 Then
Application.Calculation = xlCalculationAutomatic
ElseIf Revert = 0 Then
Application.Calculation = xlCalculationManual
End If
End Sub
This is awesome! What if you have several characters you need to have removed? What would be the easiest way as I can imagine there are several ways.?
# - 35
$ - 36
- 62
/ - 47
, - 44
. - 46
" - 34
: - 58
This is typical case of a Fitbit data export to Csv file. Each number has CHAR160 as thousand separator.. how smart Fitbit, thank you 😉
By the way, i prefer to copy the character, and use find and replace.
Sometimes it happens if you copy a table from outlook and paste it in excel. When you apply formula on those cells you will get error. What i use to do is
copy one character that looks like space,
select the entire range,
go to Find and replace,
Paste the copied character in Find option
Leave the replace option unfilled..
click on replace all..
All the errors shall be converted in to proper values..
Process looks lengthier.. but it is one of the simplest method
If Clean, Trim, and Substitute, or Find and Replace does not complete the job, I usually enter a value of 1 in an empty cell. Copy the Value of 1, Highlight the range of text numbers, and Paste Special, Values, Multiply. This site is great!
You can use Dose for Excel Add-In that can quickly clean huge data with one click besides more than +100 new functions and features to add to your Excel to save time and effort.
https://www.zbrainsoft.com
Hi,
I have a problem in excel. The sheet attached herewith.
TABLE CONFIG 2/6
A B C D E F G H
1 WEIGHT1 43,599 WEIGH2 62500 WEIGHT3 77000 WEIGHT4 66,500
2 DEDUCTION1 15,000 DEDUCTION1 15,000 TEMP 0 DEDUCTION2 11,005
3 RESULT 58,599 RESULT-1 77,500 RESULT-2 77,000 RESULT-3 77,505
4 RESULT SUBSTRACT 0 0 0
5 REQUIRED VALUE 77,500 77,000 77,505
Note: 1- RESULT (58599) IS TO BE DEDUCTION EITHER FROM D4 OR F4 OR H4 WHICHEVER IS MOST
LEAST CELL AMONG RESULT-1 OR RESULT-2 OR RESULT 3.
2-HENCE, RESULT VALUE $B$3 IS TO BE PRESENTED ON CELL EITHER D4 OR F4 OR H4 WHICHER IS
MOST LEAST VALUE
3-FORMULA =IF(E8<H8,$B$9,IF(E8<J8,$B$9,IF(H8<J8,$B$9,IF(H8<E8,$B$9,IF(J8<H8,$B$9))))))
CREATED ON CELL D4,F4 & H4 DID NOT WORK.
PLS FOR YOUR HELP.
THANK YOU
@R
Why not ask the question in the Chandoo.org Forums
https://chandoo.org/forum/
You can attach a file there