Represent a list in matrix format - excel

I have an Excel file with these columns and values:
<-To be:
I can easily insert a PivotTable and make the array but I need a formula that will add an x if a is present in 1, 3, 4, 5 and the same for b: if is present in 1 add an x, if is present in 5 add an x etc.
Is it something that can be done using an INDEX/MATCH or should a VB script be used instead?

Based on the layout from your images, you can enter this formula in cell B2 of the report sheet:
=REPT("x",0<COUNTIFS(Sheet1!$A:$A,B$1,Sheet1!$B:$B,$A2))
...and then copy over and down as far as needed.
Note: this assumes that the source list is on Sheet1.

An alternative to COUNTIFS:
=IF(SUMPRODUCT((B$1=Sheet1!$A$2:$A$10)*($A2=Sheet1!$B$2:$B$10)),"x","")

Related

Excel FILTER return top 10 while skipping columns

I have some data in a table that I would like to filter and sort the top 10 from each column but not show all columns.
Sample data:
My current formula is as follows:
=SORT(FILTER(A3:I437,C3:C437>=LARGE(C3:C437,10)),3,-1)
This outputs the following:
Ideally what I would like is to output just columns R and T from this formula but am not sure how to go about this? I will also be running the same formula to output column R with each of the other columns separately (e.g. R&S, R&U, R&V etc).
Any advice on this would be greatly appreciated!
Use the new CHOOSECOLS function if you have it:
=CHOOSECOLS(filtered_array,{1,3})
If not then you can use INDEX(filtered_array, 0, 1) to choose the entire first column, INDEX(filtered_array, 0, 3) to choose the 3rd
Alternatively to get a formula in a single cell, you can also use a second FILTER, filtering the COLS() of the array to match {1,3}, although this is a bit more involved.
N.B. Here filtered_array is your current formula
You can also CHOOSE inside the FILTER
edit: like this
Excel Filter Function - choose certain columns as output

How to Find the location of a value in a multi row/column Excel spreadsheet table?

How do I find either the cell address or preferably the row and column of the value 119, or any other number in the table below?
The table does not contain row or column titles, only the numbers as shown.
I am looking for a worksheet solution (formula) and not a VBA solution.
An Array Formula
This is an array formula and it has to be entered with control shift enter i.e. hold control shift and then press enter.
=MAX(IF(A1:J34=119,ROW(A1:J34)-ROW(A1)+1))
Remarks:
The value is searched by column i.e. A1, A2, ... B1, B2 ... i.e. if you had another 119 in cell D1 the result would still be 2, and if you had a 119 in cell c1 then the result would be
1.
For a column version just replace ROW with COLUMN:
=MAX(IF(A1:J34=119,COLUMN(A1:J34)-COLUMN(A1)+1))
Well, clunky and you can expand it, but it does work:
Row is separate to column but you could put them together in one cell, does depend on how you want to use the results, but you did not specify that so I have done this...
You could use a choose() function or a lookup table with vlookup() to change the column result to a letter...
Please try:
=MOD((K1-50),34)+1&" | "&1+(INT((K1-50)/34))
where K1 is your selected value.
Returns R | C. (Data in A1:J34 is not required.)
Below is a general purpose answer based on VBasic2008's answer.
I modified the formulas to utilize defined names so that the cell references do not have to be hard coded in the formulas. This way both the data table and row / column formulas can be relocated to anywhere on the spreadsheet. It works for both numerical AND text based data.
I also included the =ADDRESS() function to return the absolute reference of the look up value.
For illustration purposes, a step by step example for Data Set 1 is shown replacing the hard coded cell references with defined names.
The Data Set 2 section is the simplified version just using one defined for each the row and column look up value.
You can download an example spreadsheet here: Look_Up_a_Value_in_a_Table.xls
Thanks to all of you: Solar Mike, VBasic2008, and pnuts
Click on the image to enlarge.

SUMIF equivalent in array formula

I have a range of value (say H6:H20) containing values "g","y" or "r".
In order to count each status and use that as the source of a donut chart, I made a range of 3 cells calculating =COUNTIF($H$6:$H$20,"Green") etc...
Then I tried to be more elegant and create a single array formula returning the count for each letter, that I could eventually use as the source of my chart, without having intermediry calculation range. But I can't get this working.
Input:
g
r
y
g
g
r
g
g
y
Expected output (with a single array formula):
5
2
2
My try: =COUNTIF({"g","y","r"},$H$6:$H$20) -> error
Other try: =SUMPRODUCT(1*$H$6:$H$20={"g","y","r"}) -> error
(both entered in a 3 cells range with Ctrl+Shift+Enter, of course)
What is the right formula ?
You don't need an array formula, though you do need to store that formula as a Defined Name within Name Manager, e.g.:
Name: Series_Values
Refers to: =COUNTIF(Sheet1!$H$6:$H$20,{"g","y","r"})
Change the sheet name (Sheet1 here) as required.
You can then add a series to a chart with the following syntax for the Series values entry:
='Sheet1'!Series_Values
Again, amend the sheet name as required, though be sure to retain the exclamation mark and apostrophes (the latter are not strictly necessary if the worksheet name contains no spacing, though in any case it is good practice).
Excel will actually amend this to:
=Book1!Series_Values
(where Book1 is the assumed workbook name), though this is not important here.
Regards
That's what FREQUENCY function is for. Select 3 cells and CTRL+SHIFT+ENTER following formula:
=FREQUENCY(MATCH($H$6:$H$20,{"g","y","r"},0),MATCH($H$6:$H$20,{"g","y","r"},0))
Assuming this isn't dynamic, and you have a definite amount of rows, you can use the following. Let's say you're putting this formula in row 20:
=COUNTIF($A$1:$A$9,INDEX({"g","r","y"},ABS(20-(ROW()+1))))
and drag down.
Edit: If you want in the same cell, just combine COUNTIF()?
=COUNTIF($A$1:$A$9,"g")&","&COUNTIF($A$1:$A$9,"r")&","&COUNTIF($A$1:$A$9,"y")
=IF({"g","y","r"}="g",COUNTIF($H$6:$H$20,"g"),IF({"g","y","r"}="y",COUNTIF($H$6:$H$20,"y"),IF({"g","y","r"}="r",COUNTIF($H$6:$H$20,"r"))))
long solution as I don't have time to make a more elegant one. You can see the array object with F9 in the formula bar.

vlookup with multiple columns

I have the following formula in my B:B column
=VLOOKUP(A1;'mySheet'!$A:$B;2;FALSE)
It does output in B:B the values found in the mySheet!B:B where A:A = mySheet!A:A. It works fine. Now, I would like to also get the third column. It works if I add the following formula to the whole C:C column:
=VLOOKUP(A1;'mySheet'!$A:$C;3;FALSE)
However, I'm working with more than 100k lines and about 40 columns. I don't want to do 100k * 40 * VLOOKUP, I would like to only do it 100k and not have to multiply this by all the columns. Is there a way (with array-formulas maybe) to just do the VLOOKUP once per line to get all the columns I need?
data example
ID|Name
-------
1|AB
2|CB
3|DF
4|EF
ID|Column 1|Column 2
--------------------
1|somedata|whatever1
4|somedate|whatever2
3|somedaty|whatever3
I would like to get:
ID|Name|Column 1|Column 2
-------------------------
1|AB |somedata|whatever1
2|CB | |
3|DF |somedaty|whatever2
4|EF |somedate|whatever3
INDEX works fast than VLOOKUP, I would recommend using that. It'll reduce the strain that many vlookups would put on your system.
First find the row that contains what you need in a helper column with MATCH:
=MATCH(A1,'mySheet'!$A:$A,0)
Then an INDEX using that number, that you can drag across and populate all your columns with:
=INDEX('mySheet'!B:B,$B1)
Your output would be akin to:
ID|Name|Match |Column 1 |Column 2
-------------------------
1|AB |Match1|IndexCol1|IndexCol2
2|CD |Match2|IndexCol1|IndexCol2
3|EF |Match3|IndexCol1|IndexCol2
Also! I'd recomend setting these ranges to actually cover the data, rather than referencing the whole column, for additional speed gains, e.g.:
=INDEX('mySheet'!B1:B100000,$B1)
I was thinking more on your problem, and if you have contorl over the data you're looking up on, I have another suggestion you could try.
In 'mysheet', where the raw data is kept, add in a new column that concatenates each column into one cell, with some sort of unique divider not in your data:
=B1&"+"&C1&"+"&D1&"+"&E1 etc...
Then you could do one VLOOKUP or INDEX/MATCH for each row, instead of 40.
Once you have it in your new sheet, you could split the results back out.
Splitting without formulas
Copy/Paste the results of the lookup formulas as Values in the next column.
Select that column, and in the Data tab on your ribbon, select Text to Columns.
Leave it on Delimited, hit Next. Uncheck Tab, check Other, and input your delimeter (+ in my example).
Click Finish.
Splitting with formulas
Use =FIND() to locate each delimter, and =MID() to pull out the text between each set of delimeters, using the previous delimeter as the Start_num.
Definitely the more complex of the two methods.
If I'm understanding correctly one thing I would do to start would be to use =VLOOKUP(A1;'mySheet'!$A:LastColumn;COLUMN(B1);FALSE). This way your column reference will move as you drag your Vlookup to the right.
No formula.No output. So there can't be a way to apply formula on 1 column only and get on the others.
The other feasible way is, put i formula in 1 cell, use $ signs inteligently and drag across all cells in a giffy without having to put vlookup 40 times.
Vlookup has 4 codes to input
1-Lookup Value. Use this $A1 (put $ on A and not 1)
2-Source data- Put $ signs everywhere
3-Column index no. Just above your entire data,in the 1st row,add an empty row.Put the values 1 in A1, 2 in B1, 3 in C1 and so on. Now in the formula,instead of manually putting "2" or "3" Give reference to these cells.Put $ on Numberal and not column ( B$1).
4- Type false or 0
Then drag this across everywhere.
Lookup Value. Use this $A1 (put $ on A and not 1)
Source data- Put $ signs everywhere
Column index no. Just use column name from where data needs to be pulled (e.g. COLUMN(B1) if Lookup value is in Column A and you want value from column B).
Type false or 0

How to use a named column in Excel formulas

I know how to make a named range in Excel.
I have a spreadsheet, with various columns going across as parameters, and then finally a formula in the last cell. This is repeated many times in each row, with each row having a different set of data, and the formula updated to reference the correct row index.
However, the formula looks like (three rows worth):
=G2*(10*D2 + 20*E2 + 5*F2)
=G3*(10*D3 + 20*E3 + 5*F3)
=G4*(10*D4 + 20*E4 + 5*F4)
I would like to use named ranges, but I can't find a way to do something like
=Count * (10*var1 + 20*var2 + 5*var3)
where count, var1, var2, and var3 automatically update to be the particular column of the current row. I can create a named range for every cell, but that isn't helpful. I can name range the column, but then I can't find a way to put an offset into the formula.
Also the whole point of this is readability, so if it ends up being some nasty complex formula function call, that probably doesn't help too much.
Simple, at least when using Excel 2010:
name your column: select full column, enter name
use column name in formula; Excel will combine the referenced column with the current row to access a single cell.
Using the example from Alex P:
select column D by clicking the column header containing the "D", enter name "input1" into name field, and press Enter.
repeat for columns E to F, using "input2" and "input3", respectively.
Do not define additional names defining names "input1" [...] as in example above!
use the formula as given in the example above
Attention:
Using named columns this way, you cannot access any other row as the one your formula is in!
At least I'm not aware of the possibility to express something like <ColName>(row+1)...
I would suggest creating a Table. Select your range A1:H4, then go to the Tables widget > New > Insert Table with Headers (on Mac). This will mark A2:H4 as body of the table, and A1:H4 as header.
From that, you get:
Whatever you put into the header column will define the name for this column automatically, e.g. Count, Radius, Density, Height
You can then write your formula using =[#Count]*(10*[#Radius] + 20*[#Density] + 5*[#Height])
When you change the formula in cell H2, Excel will automatically "copy down" this formula to all cells in column H. So no more accidental inconsistencies in the formulas.
When you need to add another row, simply click the last cell (in our example H4) and hit Tab. Excel adds another row, and also makes sure to "copy down" your formula into the new row.
If you need a total row, add it with the Total Row checkbox in the Tables widget. Excel adds a total row automatically. If you click any cell in the total row, you can change the "total formula" with the "▼▲" button, for example to calculate the Average instead of the Sum of the column.
If you have a long table and scroll down so that the header is not visible anymore, Excel automatically displays the column header instead of the column names (Count instead of G for example).
I can really recommend the video You Suck at Excel with Joel Spolsky which explains all of that.
Suppose I have the following numbers set up in columns D to F in rows 2 to 4:
D E F G
2 10 15 20
3 1 2 3
4 20 30 40
Now suppose I want the value in column D to be known as input1, column E to be input2, and column F to input3:
In Insert > Name > Define...
input1 RefersTo =OFFSET(Sheet1!$D$2,0,0,COUNT(Sheet1!$D:$D),1)
input2 RefersTo =OFFSET(Sheet1!$E$2,0,0,COUNT(Sheet1!$E:$E),1)
input3 RefersTo =OFFSET(Sheet1!$F$2,0,0,COUNT(Sheet1!$F:$F),1)
Now if I write my formula in column G as follows I should get correct answers:
G2 =(10*input1+20*input2+30*input3) // 1000
G3 =(10*input1+20*input2+30*input3) // 140
G5 =(10*input1+20*input2+30*input3) // 2000
I haven't fully reviewed the previous answers, but I think this is closer to what #Jason Coyne the OP was looking for. So, I hope I get a lot of up votes. ;-)
Excel allows your formula to refer to tables and columns by name if you "Format as Table". Here is an article titled Using structured references with Excel tables that goes into detail.
FWIW, it looks like this feature has been available since Excel 2007.
Here is a screenshot of an example:
You should be able to see the formula in E2 is =[#Count] * (10*[#Var1] + 20*[#Var2] + 5*[#Var3]) which is pretty close to what #jason-coyne wanted to type.
I don't like that you are forced to pick a style (or define a new one if you don't see a style you like). The good news is you can reformat the cells all you wish without undoing the "tableness".
It insists on turning on auto-filter. But, auto filter is easy to turn off (see the Filter Button checkbox under the Table Tools Design menu).
It also insists on having non-empty, unique values in the header row (Which kinda makes sense). If you delete a header cell, or insert a column, Excel will invent a new, unique name and stuff it in for you. D'oh!
If you want a column to not have a header, you can enter an apostrophe (') followed by one or more blanks. Remember header values need to be unique, so keep adding blanks if you want more than one column without a header.
If you would like to download the sample workbook in the screenshot, here is a link: https://filebin.ca/3vfaSDn4NLEA/SampleWorkbook.xlsx
Adding to Alex P's answer:
Instead of using =OFFSET(Sheet1!$D$2,0,0,COUNT(Sheet1!$D:$D),1) as the formula for input1, I recommend to use =Sheet1!$D$2:INDEX(Sheet1!$D:$D,COUNT(Sheet1"$D:$D))
It produces the same result, but it is non-volatile, i.e., only recalculate when a predecessor cell changes. This is much better in a larger model!
If you're using VBA, then you can select the whole column and name it, say MyCol, in the name box (upper left input box). The in your code you can refer to a cell in the column MyCol (line 12) using the following code:
Cells(12, Range("MyCol").Column)
You might be able to use the row() function. This returns the current row that you are in. So depending on the layout of the spreadsheet you can use it like this:
=offset(NamedColumn1, row()-1)
The -1 is because you are saying how many rows to move down from row 1 which if you are in row 1 you want to be 0.
Use the Excel feature called named references.
To name a cell or range of cells
select that cell or range of cells
Enter its name in the Name Box ( its left of the formula widget and has the cell name )
You can't use names that conflict with cell names, like k0.
The named cells can be used if formulas. E.g.,
=pi*radius*radius
I'd like to propose a slight variation of the cell reference made by Dror. This will work as well:
Range("MyCol").Rows(12)

Resources