Am running Excel 2013 on Windows 10 with 16 GB memory.
I have a hierarchical structure similar to figure 1 below. My structure is 1000+ rows and as deep as 27 levels.
For a "tree" as depicted above, how to identify the parent of a cell?
Here's what I've done so far. Using conditional formatting, and this formula:
=AND(A1<>"",OFFSET(A1,1,1)="")
I am able to identify the leaf nodes of the tree, as shown in figure 2 below.
Now I need to find the immediate ancestor (parent) of every cell (child) in the tree (with the exception of the first row that does not have a parent).
Figure 3 (below) shows an example done manually:
Further, I can associate the tree with an outline.
Then I am able to count the number of decimals in the outline in order to track level changes using this formula:
=LEN(A2)-LEN(SUBSTITUTE(A2,".",""))
as shown in figure 4 below:
How to programmatically accomplish what was done manually in figure 3?
For each row (recorded in a column outside of the tree), print the location of the cell that is the parent of the cell.
Here is a similar question that manually provides "level" information.
How to find parent in an indented hierarchy?
In my case, I need to work from an outline (as shown) or by using a formula for Conditional formatting.
According to your rules, the parent of a cell is the first non-empty cell, bottom to top, in the column immediately to the left:
public function GetParentCell(byval c as range) as range
set GetParentCell = c.offset(0,-1).end(xlUp)
end function
Having put this in a module, you can use it from the sheet, e.g.
=CELL("address", GetParentCell(D4))
to display the address of the parent of D4.
If you want to avoid VBA completely, you have to find a way to convert End(xlUp) into a formula which may prove difficult because Excel lookups work from top to bottom and you want the other way round. If I was to do that, I would start from Last non-empty cell in a column.
Given that the hierarchy is contained in A1:G19:
In H1, put:
=MATCH("x",$A1:$G1,0)
In I1, put
=CELL("address",INDEX(INDEX($A$1:$G$19,1,H1-1):INDEX($A$1:$G$19,ROW(),H1-1),MAX((INDEX($A$1:$G$19,1,H1-1):INDEX($A$1:$G$19,ROW(),H1-1)<>"")*(ROW(INDEX($A$1:$G$19,1,H1-1):INDEX($A$1:$G$19,ROW(),H1-1))))))
as an array formula (CTRL+Shift+Enter).
The credit goes to https://stackoverflow.com/a/5442955/11683 because it is nothing more than a dynamic version of =INDEX(A:A,MAX((A:A<>"")*(ROW(A:A)))).
Drag everything down.
The formula in column I returns the parent of each cell as a true reference which is then examined by the CELL() function to show the address. You can remove the CELL() and use the calculated reference in a different way.
Related
So I'm not a fan of VBA and I recently learned that OFFSET can be used with COUNTA to flashfill a range as far at it is as long as you aim for a longer range than you have data. Now I want to be able to achieve this both for columns and rows at the same time, where the rows are averaged. Could this be done? I am banging my head against the wall to find some logic to do it, but can only manage to combine it in a way that multiplies the rows with the number of the column.. which is not desired, of course.
I have posted a Minimal Reproducible Example in Excel Online:
https://onedrive.live.com/view.aspx?resid=63EC0594BD919535!1491&ithint=file%2cxlsx&authkey=!ALmV0VtFb7QZCvI
If you see Cell J9 and J11 you will see what I want to combine. The three rows in J11 and down, I want to average in J10, and spill/flashfill (like J9 and 11 does automatically because of the formula already) them from to the right, for as many columns as there data in the range A1-G4..
So I have raw data of numbers with titles in A1-G4, and by writing =OFFSET($A$1:$A$1,0,0,1,COUNTA($A$1:$EV$1)-1) in J9 I get all the titles of the columns filled from left to right, and by writing =OFFSET($A$1,1,0,COUNTA($A:$A)-1) in J11 I get the rows of the first column filled from top to bottom. They can also be combined, by writing OFFSET(Days,1,0,COUNTA($A:$A)-1,COUNTA(Days)), where "Days" is =OFFSET($A$1:$A$1,0,0,1,COUNTA($A$1:$EV$1)-1) (in a named range for readability) or OFFSET($A$1:$A$1,0,0,1,COUNTA($A$1:$EV$1)-1) without using a named range
As a thought, though I'm not sure how to implement it, maybe this could somehow be used in some form to get the column reference for the horizontal part in combination with =AVERAGE(OFFSET($A$1,1,0,COUNTA($A:$A)-1))
=MID(ADDRESS(ROW(),COLUMN()),2,SEARCH("$",ADDRESS(ROW(),COLUMN()),2)-2)
..found at https://superuser.com/questions/1259506/formula-to-return-just-the-column-letter-in-excel/1259507
Now, based on your explanation, here is the screenshot of my test:
Section A1:Exxx
I have converted that section into a Table, called «TblData», having numerous avantages:
It expands automatically without any additional efforts/formula
We can identify Data by its Columns attributed automatically by the Table [#1], [#2],[#3], [#4], [#5]
Section J9:N9
As a replica of the table name, I have used the following formula to retrieve it:
=INDEX(TblData[#Headers],1,COLUMN(A1)) '<--- This is for J9
=INDEX(TblData[#Headers],1,COLUMN(E1)) '<--- This is for N9
Section J11:Nxx
As a replica of the Table Content, I have used the following formula to populate the content:
=INDEX(TblData,ROW($A1),MATCH(J$9,TblData[#Headers],0)) '<--- This is on J11
=INDEX(TblData,ROW($A3),MATCH(N$9,TblData[#Headers],0)) '<--- This is on N13
Section J10:N10
Now this is the interesting part of the Average, so here is the formula I used for it:
=AVERAGE(TblData[1]) '<--- This is on J10
=AVERAGE(TblData[5]) '<--- This is on N10
NB: (1) Instead of using the Content below J10:N10, I prefer to reuse the Table as it expands automatically as more rows are added.
(2) Unless it is really necessary, I feel it is a double work as well to replicate again A1:Exxx from J9:Nxxx, because you can use the Table for whatever you need, with less maintenance.
Kindly find attached the file as well after I updated those items:
File Link: https://drive.google.com/open?id=1wRbpUxg0XLpfGqdvMF4fNKXDrL7xPPWs
We can correspond more below for further info. Hoping you to strech more your compentence :)
Sorry, mate, I can't figure out what you want to calculate. If it makes sense to add J9+J11 then you could just concatenate the two formulas in J9 and J11 with a plus sign. After much deliberation I decided to assume that your question is not one of formula but of formula-writing - "referencing" for short. Therefore I prepared this answer for you, hoping that it will prove helpful.
Building on your named range Days I suggest you create a dynamic named range Data with this formula.
[Data] =OFFSET(Sheet1!$A$1,0,0,COUNTA(Sheet1!$A:$A),COUNTA(Sheet1!$1:$1))
The range thus defined is dynamic in both directions. However, bearing in mind that OFFSET is volatile (slows down your worksheet) you may like to keep its use limited to this one formula and perhaps start the range at A2, but I shall tempt you to break the rule. Now you can use the INDEX function to refer to the Data range.
= INDEX(Data, [Row number], [Column number])
defines a single cell. But by setting either column or row to zero you can define an entire column or row. =INDEX(Data,0,1) defines column 1 of the Data range, =INDEX(Data,1,0) defines its first row.
=INDEX(OFFSET(Data,1,0),0,1) defines the first column of a range moved down by one row from its original position. I recommend the alternative and start the Data range from A2 and perhaps declare another range for the first row if needed.
=AVERAGE(INDEX(Data,0,1)) would draw the same average you already have in your sheet, provided that Data was defined starting at A2. For fun's sake, =AVERAGE(INDEX(OFFSET(Data,1,0),0,1)) would do the same without the change in the range's definition.
=COLUMN() returns the number of the column this formula resides in. So, you could enter =COLUMN()-6 in column G, copy to the right and get a count starting from 1. (You can do the same vertically with the ROW() function.) Applied to your formula, =AVERAGE(INDEX(Data,0,COLUMN()-6)) would return the average from column 1 if entered in column G, and from columns 2, 3 4, etc as copied to the right.
As I said, I don't understand enough of your request to bring this idea to a conclusion but I think that using the method described above will provide you with a tool to copy formulas into the table your sample has at its right. If you would elaborate on your requirement I might be able to assist more.
I have been trying to make a form for some of my team members who are not that computer literate, and I essentially want to make it click and go. I thought I could do it...but alas I am not as good with nesting functions as I thought I was.
I have this spreadsheet where I want to put data into the yellow cell. On the next sheet I have the below table. What I want to do is use a formula to fill H4 with the "Request Branch's" Account Number. Now, I have currently filled the cells with information. They, in fact, have drop down options - which are pulled from the Account List table. As a result the value in H4 will continually change based on the needs of the user - but must be within the confines of the Account List Table.
What I have tried is here and enter link description here. I keep getting result of #Value, or N/A. I can't figure out what I am doing wrong. I know that I need to nest the SUMIFS withing VLookUp, but I am not sure as to why it won't work.
I'm providing you with two possible solutions.
1) The first one uses the SUMPRODUCT function. You may not have seen this kind of notation before.
When ranges are multiplied by each other like so (B3:B8=G3)*(C3:C8=G4) they are actually turned into boolean arguments. If you highlighted this part of the code and pressed F9 it would look like this: {0;0;0;0;1;0;0}. This is an array where TRUE for both criteria meet. So our Branch is "A" and our Carrier is "F". In the rest of the cases either or both are false resulting in zeroes.
Now if you multiply this array by the range with account numbers, obviously the only number remaining will be the one multiplied by 1 and so you have the answer however keep in mind that as you are multiplying if the account is not a number the function will fail!
2) This is why we have a second method using =INDEX() and =MATCH() functions.
To overly simplify this - the INDEX function grabs contents from an array at a specified position (row and column), while the MATCH function gets the position of an item in an array.
The idea with using ranges as multiple criteria is the same as in the first example, however this time when we get our array of zeroes and ones {0;0;0;0;1;0;0} we use the match function to find at which position our criteria cross (as seen on the screenshot it's the 5th position, as it's in the 5th row of the entire column D, the match function searches the {0;0;0;0;1;0;0} array for a 1 and returns its position in the array) and so this is our ROW.
Knowing the position of the contents we searched for we use the INDEX function to grab the contents of the cell in that position so =INDEX(D2:D8,MATCH(1, INDEX((B2:B8=G3)*(C2:C8=G4),0),0)) is actually =INDEX(D2:D8, 5) meaning that the INDEX function grabs contents of the 5th row from the range D2:D8 which is cell D6.
The green boxes are just there to show the instance where both of our criteria are met (cross).
Please try this formula.
=INDEX(Table1[Account],SUMPRODUCT((Table1[Branch]=F$3)*(Table1[Carrier]=F$4)*ROW(Table1[Account]))-1)
Note that you may have to adjust the final -1. This is because the indexed table starts in row #2 and the SUMPRODUCT function returns a sheet row number. If your table would start in a different row the difference between the sheet row and table row would be larger. It must be adjusted here. Or you might work with sheet references (named ranges) and require no adjustment at all.
I have a data in Excel as shown in attached image where in I've used named ranges.
Spends. Apr18. May18. Jun18
Category A. 120. 120. 120
Category B. 135. 125. 129
Category C. 110. 111. 112
Name Range: Spends
Range: =Sheet1!$A$1:$D$4
The average quarterly spends are calculated using named references as:
=AVERAGE(INDEX(Spends,2,2):INDEX(Spends,2,4))
This returns 119 in Cell E2
How can I drag this formula to subsequent cells so that it is applied automatically.
If you use the INDEX function with constants, of course these constants will not change when you drag the formula down and/or across.
You will need to learn about absolute and relative references and use something like Row(A1) and/or Column(A1) instead of the constants 2 and 2.
But maybe it does not have to be that complicated.
You could just add a column in the data table that calculates the Average and refer to that cell.
Or, in cell E2, use the formula =average(B2:D2).
I don't see the need for the named range at all in this scenario. It only complicates things.
It seems that you are struggling with appropriate data architecture concepts rather than with using formulas.
If in doubt, keep it simple. I don't see an application for copying the formula across, since you only have three months of data and you want to average these three months.
If your real scenario has more columns, then, please!!, update your question and post more relevant context. Then post a comment, so I can see you made a change.
Again, since you are new here: do not post updates into comments. Edit your question and then post a comment.
Use the ROW() function to return the relative position
E.g. in E2 and drag down
=AVERAGE(INDEX(Spends,ROW(),2):INDEX(Spends,ROW(),4))
Maybe you want to dynamically find the quarter start and end columns then you can use Match function to find the dates and return the position (column) where found. Then feed these into your formula:
Getting the columns by searching for qtr start and end:
Referencing those found positions as column arguments in your formula:
I've been struggling with this longer than I care to admit, but I have a fairly simple OFFSET function call which works on one sheet, but if I copy it to a different sheet it gives a #VALUE error.
On a sheet named "Deliverable" I have this formula in a cell:
=OFFSET(Deliverable!$B$72,1,0,,3)
and it works fine.
If I go to any other sheet and use the same exact formula, or use it in the Name Manager, it gives a #VALUE error.
If I leave off the final parameter indicated the number of columns I want, it does work:
=OFFSET(Deliverable!$B$72,1,0)
but of course isn't giving me the range I need.
Any idea what's going on with this?
I'm using Excel 2016 on Windows 7.
-- Updated Info --
In a nutshell, my spreadsheet has two cells which I'm using as dropdown lists, where the 2nd cell's list feeds off the selection in the first. The data they are based on has this format:
OptionA A B C D
OptionB A B
OptionC D E F
So the first dropdown uses a simple Data Validation source pointing to the column with OptionA, OptionB, etc. Once that's chosen, the second dropdown list should contain the appropriate options for the one selected. So if OptionB is selected, then the 2nd dropdown list should show A and B.
When I initially wrote this, the data validation source was just a simple VLOOKUP entry, but the lists often had blanks since the number of options varies for each entry. Wanting to fix it up a bit, I ended up with this formula:
=OFFSET(Deliverable!B72,Deliverable!B87,0,1,COUNTA(OFFSET(Deliverable!B72,Deliverable!B87,0,1,5)))
There won't be any more than 5 options, and there are no empty cells in the middle of the data to filter out.
In one spreadsheet I have I used this as a named range definition, then specified the named range for the cells data validation source and it worked. In this other spreadsheet however, it gave me the error described earlier.
However, it looks like when I enter the statement directly into the data validation source field and not in the name manager, it works as expected.
Am I taking the totally wrong approach?
What is it that you want this formula to do? As written, it is returning a block of three horizontal cells. The #VALUE error is Excel's way of telling you "Hey, you're trying to return three cells, but I can't fit them all in the one cell that you are calling this formula from".
The reason you see a result in some places and not others is because of something called Implicit Intersection. Give it a spin on Google. But basically, it just returns whichever one of those three results corresponds to the column that the formula is entered into. If you copy that exact same formula to say row F you will see that it returns a #VALUE error there, because it doesn't know what cell it should return given the column you're calling it from doesn't match any of the cells it is returning. The fact that you don't know this indicates that the formula you're using doesn't in fact do what you think it does.
--UPDATE --
Okay, following your further clarificaiton it seems that you're talking about Cascading Dropdowns aka Dynamic Dropdowns. Lots of info on Google about how to set these up, but you may be interested in an approach I blogged about sometime back that not only provides this functionality, but also ensures that someone can't later on go and change the 'upstream' dropdown without first clearing the 'downstream' one should they want to make a change.
Note that those links talk about a slightly complicated method compared to others, but the method has it's advantages in that it also handles more levels than two, and your DV lists are easily maintained as they live in an Excel Table.
This sounds like an array equation. Try hitting Ctrl+Shift+Enter in the other sheets to validate it as an array equation.
Whenever you need to reference ranges instead of single cells, Excel needs to know that you are working with arrays.
Need solutions for the queries posted in the below link,
http://www.mrexcel.com/forum/excel-questions/334740-dynamically-sum-row-values-based-column-headers.html
Having PersonWeek as separate column with sum of weeks should be displayed.
Quest: Dynamically sum row values based on column headers?
It is good form to re-post the question in its entirety on this site; nevertheless:
What you are looking for is relatively straight forward. There are a lot of ways to accomplish this but I suggest that you use combination of the 'Match' function and the the 'Offset' function. The Offset function creates a range based on a given start point, moving up/down / left/right as indicated, with a given height / width. For example:
=Offset(B2,1,2,3,4)
Indicates the range D3:G5. This is the range given by starting at cell B2, moving 1 row down and 2 columns to the right, and going for a total of 3 rows and 4 columns.
So the remaining point is to determine where to start and stop your offset.
First, the first field in your Offset function will be cell A1, as that is the top-left corner of your data table. To find how many rows to move down, you need to find what project you are referring to (I will assume that cell A6 is where you enter the project name that you care about, A7 is where you enter the first week you care about, and A8 is where you enter the last week you care about). To find how many rows to move down from A1, then, use Match:
=Match(A6, A2:A5,0)
To find how many columns to move to the right to find your first week, use Match again:
=Match(A7, B1:G1,0)
Assuming you only want to look at a single project, we know how high we want the range to be (1).
To find how wide we want the range to be, we need to know your ending week, less your starting week:
=(Match(A8,B1:G1,0)-Match(A7, B1:G1,0))
So the whole thing together will be:
=Offset(A1,Match(A6, A2:A5,0),Match(A7, B1:G1,0),1,Match(A8,B1:G1,0)-Match(A7, B1:G1,0))
Now the only thing left is to wrap the newly defined range in a 'sum' function, like so:
=Sum(Offset(A1,Match(A6, A2:A5,0),Match(A7, B1:G1,0),1,Match(A8,B1:G1,0)-Match(A7, B1:G1,0)))
This formula will create an error if someone enters a project name / weekname in a way that isn't found in your table - so in cells A6-A8 you may want to use data validation to only allow those names to be entered - let me know if you would like elaboration on that.