I have built a large data set and I need to see the average results given many different criteria. I've done this with the AVERAGEIFS function and it works just fine, however the more and more I add its getting really time intensive.
I'm wondering if there is a way to nest a vlookup or index match or anything like that in the AVERAGEIFS that read the criteria column heading and criteria in a cell (or 2 if they need to be separated) to be added to the AVERAGEIFS.
Here is an example of my spreadsheet:
The first 3 sets of criteria I want to stay locked.
I want it to read what the 4th criteria column and criteria should be by referencing the I11 cell. The highlighted portion in the formula bar is the part that I want to reference I11 so it reads it and knows that the 4th criteria is the 'code' column and the criteria is '>7'. I can separate this into 2 separate cells if need be.
I've tried a few combinations of VLOOKUP and INDEX MATCH but cannot get it to work.
Data as Text:
Price,Type,sub cat,Time,code,amount,Result,,
,,,,,,,,
9.95,t2,d,ac,2.18," 22,780,893 ",0.73,,T2 and D and AC
118.94,u2,d,bo,2.78," 172,110,893 ",4.07,,
57.63,t1,u,ac,7.09," 128,419,877 ",-2.16,,code
8.88,t2,d,ac,1.50," 62,634,868 ",12.72,,amount < 100 000 000
11.61,u1,u,ac,2.14," 146,982,736 ",1.07,,price >10
13.46,u3,u,ac,0.93," 17,513,672 ",-13.93,,
31.53,t1,u,ac,0.89," 47,170,877 ",1.39,,
16.34,t3,d,bo,1.07," 1,914,767,076 ",-1.42,,
111.59,u1,d,bo,0.62," 2,283,546,000 ",0.67,,
72.4,u3,d,bo,10.37," 951,541,514 ",1.13,,
34.55,u3,d,bo,0.77," 951,541,514 ",-2.52,,
42.25,t1,d,bo,1.05," 63,748,352 ",8.88,,
17.18,u3,u,ac,2.64," 140,217,257 ",4.35,,
97.66,t1,d,bo,3.45," 1,070,383,954 ",1.33,,
58.49,t2,u,bo,8.64," 151,876,559 ",-0.92,,
64.48,t2,d,ac,2.35," 291,967,334 ",3.03,,
38.4,t1,u,ac,17.05," 83,478,472 ",-4.31,,
20.87,u3,d,ac,28.92," 214,080,937 ",-2.16,,
36.53,t1,d,ac,1.43," 73,438,589 ",-2.07,,
89.16,t3,u,ac,1.41," 26,786,958 ",-1.75,,
15.84,t1,u,bo,2.90," 133,560,818 ",1.76,,
3.2,u3,u,bo,2.95," 215,677,667 ",-1.06,,
25.46,t1,d,bo,3.92," 57,148,431 ",1.89,,
40,t2,d,ac,8.00," 65,274,903 ",0.61,,
27.72,t1,u,ac,2.50," 381,400,886 ",6.46,,
29.07,u3,u,ac,2.32," 52,632,107 ",-0.78,,
173.31,t1,d,ac,3.58," 31,547,380 ",-4.92,,
18.22,u3,d,ac,0.58," 292,669,493 ",4.06,,
9.59,t1,d,bo,3.60," 266,883,020 ",3.16,,
115.22,t2,d,bo,4.51," 132,376,476 ",0.78,,
64.48,u3,d,ac,3.03," 338,360,104 ",-0.95,,
41.74,t1,u,bo,25.65," 245,766,436 ",-3.42,,
5.99,t3,u,bo,2.15," 175,054,713 ",-4.37,,
Use INDEX/MATCH to return the correct column. This will require that you separate the column name and the criteria:
=AVERAGEIFS(G:G,B:B,"T2",C:C,"D",D:D,"AC",INDEX(A:F,0,MATCH(I11,$A$7:$G$7,0)),J11)
An idea:
I10 - "Write down the limitation. (You have to use <,>,=,<> AND the value, for e.g.: <5)"
I11 - The user can use relations and values.
In J11, you can reference to I11 ;) It works for me.
Working with 2 separate data sets (with duplicates)
Dataset is unique identified by an ID.
There may not be an entry for the timestamp I require.
Datasets are quite large, and due to duplicates, can't use vlookup.
Samples:
Table 1:
Device Name|Time Bracket| On/Off?
ID1 |06:20:00 |
ID2 |06:20:00 |
ID3 |06:30:00 |
Table 2:
Device Name |Timestamp |On/Off?
ID1 |06:20:00 |On
ID2 |06:50:00 |Off
ID3 |07:20:00 |Off
What I want to achieve:
I want an if statement to check if:
1) device ID matches AND
2) timestamp matches
If so, return the value of On/Off from Table 2.
If not, then I want it to return the value of the cell above it IF it's the same device, otherwise just put "absent" into the cell.
I thought I could do this with some IF statements like so:
=if(HOUR([#[Time Bracket]]) = HOUR(Table13[#[Timestamp Rounded (GMT)]]) and
minute([#[Time Bracket]]) = minute(Table13[#[Timestamp Rounded (GMT)]]) and
[#[Device Name]]=Table13[#[Device Name]], Table13[#[On/Off?]],
IF([#[Device Name]]=Table13[#[Device Name]], INDIRECT("B" and Rows()-1), "absent"))
(I put some newlines in there for readability)
However, this doesn't seem to resolve at all... what am I doing wrong?
Is this even the correct way of achieving this?
I've also tried something similar with a VLookUp, but that failed horribly.
Thanks all!
To not deal with array formulas or merging strings which, (not in your case) can still be wrong at the end, I suggest the use of COUNTIFS due to the fact, you have a very small amount of outcomes (just on or off)...
for the first table (starting at A1, so the formula is at C2):
=IFERROR(CHOOSE(
OR(COUNTIFS(Table13[Device Name],[#[Device Name]],Table13[Timestamp],[#[Time Bracket]],Table13[On/Off?],"On"))+
OR(COUNTIFS(Table13[Device Name],[#[Device Name]],Table13[Timestamp],[#[Time Bracket]],Table13[On/Off?],"Off"))*2
,"On","Off","Error"),IF(A1=[#[Device Name]],C1,"Absent"))
this will also show "Error" of a match for "On" and "Off" is shown... to skip that and increase the speed, you also could use:
=IF(COUNTIFS(Table13[Device Name],[#[Device Name]],Table13[Timestamp],[#[Time Bracket]],Table13[On/Off?],"On"),"On",
IF(COUNTIFS(Table13[Device Name],[#[Device Name]],Table13[Timestamp],[#[Time Bracket]],Table13[On/Off?],"Off"),"Off",
IF(A1=[#[Device Name]],C1,"Absent")))
For both the "Device Name" is at column A, "Time Bracket" at column B and "On/Off?" at column C while the table starts at row 1... If that is not the case for you, then change A1 and C1 so they match
(Also inserted line-breaks for better reading)
Picture to show the layout:
I picked the second formula to show how it works... also, this formula should not be able to return 0's... I'm confused
Couple of good suggestions, however using the helper column as suggested in the topic by Scott Craner above worked.
Created a helper column of concat'd device ID and timestamp for both tables, then did a simple VlookUp.
Another lesson learned: Think outside of the box, and go with simple solutions, rather than try + be too clever like I was doing... :)
I'm inexperienced in Excel functions and need some help with the following scenario:
Table I has 2 columns:
X Y
3.2 result
4.7 result
1.2 result
Table II has these columns:
A B C D
1 1.2 0.0 2.3
2 4.1 3.2 0.0
3 0.0 4.7 0.0
I will try to illustrate what I want to do in C++/C#/Java programming syntax:
In Y[1] I want to store:
foreach (int value in TableII) //look for X[1] in Table 2
if( value == X[1] ){ //when you find it,
return A[value.rowNumber]; //return the corresponding value from A
//break;
}
In other words, I want to find X[1] in table II (it will always be there, and always just once), and return the corresponding value from column A[of the same index as the match].
I've looked at HLOOKUP but there's more to it that just one command. Please help.
[EDIT]
I've tried this =INDEX(Column1,IFERROR(MATCH(I2,Column2,0),IFERROR(MATCH(I2,Column3,0),IFERROR(MATCH(I2,Column4,0),IFERROR(MATCH(I2,Column5,0),IFERROR(MATCH(I2,Column6,0),IFERROR(MATCH(I2,Column7,0),IFERROR(MATCH(I2,Column8,0),MATCH(I2,Column9,0)))))))))
But I get the "more levels of nesting than are allowed in the current file format" error. If I remove just one IFERROR(MATCH(I2,Column8,0), then it works. But I need Columns 2 to 9, with Column1 holding the index.
I thought Excel 2007 and 2010 do not have nesting limits any more. What gives?
(For anyone new to Excel reading this, "Column1" etc. are made by selecting the column, right clicking on the selection, and "Define Name".)
[EDIT2]
So I "solved" my nesting problem by dividing the formula into 2 parts, with a +, having the ifs return 0 in case there's no match in either side. This way you either have something + 0 or 0 + something, so it works.
Here's the formula I'm using right now:
=INDEX(Column1, IFERROR(MATCH(I3,Column2,0),IFERROR(MATCH(I3,Column3,0),IFERROR(MATCH(I3,Column4,0),IFERROR(MATCH(I3,Column5,0),IFERROR(MATCH(I3,Column6,0),IFERROR(MATCH(I3,Column7,0),0))))))
+
IFERROR(MATCH(I3,Column8,0),IFERROR(MATCH(I3,Column9,0),0))
)
I've accepted Chris' answer, as he did most of the work, but I would still like to know if my workaround is good practice and if there isn't a shorter/better way to do this.
Thank you.
Here's a possible formula for Y (assumes table I is located in columns G:H)
=INDEX($A:$A,IFERROR(MATCH(G2,$B:$B,0),IFERROR(MATCH(G2,$C:$C,0),MATCH(G2,$D:$D,0))))
Note the IsError function was introduced in Ecel 2007
If your tables are actually defined as excel tables you can use
=INDEX(Table2[Column1],IFERROR(MATCH([#X],Table2[Column2],0),IFERROR(MATCH([#X],Table2[Column3],0),MATCH([#X],Table2[Column4],0))))
(substitute your own table and column names)
If your values are unique, you can use the following formula without having to do any nesting:
=INDEX(column1,SUMPRODUCT((Table2=G24)*ROW(Table2)))