SOLVED

Need help to resolve Excel Formula using Let,Take,index,Sort,Unique,HSTACK,MMULT,TOROW,MONTH.

Brass Contributor

I am trying to add 5 additional columns to the function but can get only a $value:

 

Function :

 

=LET(r,B1:G8,m,6,a,TAKE(r,,1),b,INDEX(r,,2),c,INDEX(r,,3),d,INDEX(r,,4),e,INDEX(r,,5),f,INDEX(r,,6),g,INDEX(r,7),u,SORT(UNIQUE(a)),HSTACK(u,MMULT(--(TOROW(a)=u),(MONTH(b)=m)*c)))

 

Data
A                       B                      C           D          E                                     F                                     G
Chequing 06962-5001375 2023-08-01 " " INSURANCE IND ALL AC-EST                                  -203.64
Chequing 06962-5001375 2023-08-01 " " INSURANCE THE PERSONAL                                   -269.88
Chequing 06962-5001375 2023-08-02 " " OD HANDLING FEE 1 @ $ 5.00                                        -5
Chequing 06962-5001375 2023-08-09 " " Transfer WWW TRANSFER - 0982                           1000.00
Chequing 06962-5001375 2023-08-09 " " Payment WWW PAYMENT - 5925 MILTON - TAXE         -50
Chequing 06962-5001375 2023-08-09 " " Payment WWW PAYMENT - 2317 RELIANCE                  -55
Chequing 06962-5001375 2023-08-09 " " Payment WWW PAYMENT - 8635 MILTON HYDRO -145.52
Chequing 06962-5001367 2023-04-18 OD INTEREST -3.7

 

Result #VALUE

 

If I use a subscript of the data i.e. 3 col The function is of:

 

Function:

=LET(r,A1:G8,m,6,a,TAKE(r,,1),b,INDEX(r,,2),c,INDEX(r,,3),u,SORT(UNIQUE(a)),HSTACK(u,MMULT(--(TOROW(a)=u),(MONTH(b)=m)*c)))

 

 

Data:

 

06962-5001375 2013-06-10 2.00
06962-5001375 2013-06-10 3.00
06962-5001375 2013-05-13 4.00
06962-5001375 2013-05-14 5.00
06962-5001375 2013-06-15 6.00
06962-5001380 2013-06-17 7.00
03092-5224710 2013-06-17 8.00
03092-5224710 2013-06-18 9.00


Result:

06962-50011121 3
06962-50011122 3
06962-50011123 6

Thanks tp P. b. at stack overflow for the original suggestion.


How do I add in the extra columns?

 

I would appreciate any help that you can provide me

 

Thanking you in advance.

86 Replies

@Norman_McIlwaine 

In general your formula

=LET(
    r, B1:G8,
    m, 8,
    a, TAKE(r, , 1),
    b, INDEX(r, , 2),
    c, INDEX(r, , 3),
    d, INDEX(r, , 4),
    e, INDEX(r, , 5),
    f, INDEX(r, , 6),
    g, INDEX(r, 7),
    u, SORT(UNIQUE(a)),
    HSTACK(u, MMULT(--(TOROW(a) = u), (MONTH(b) = m) * c))
)

works, e.g.

image.png

but I was not able to paste your sample into the grid and didn't catch what exactly you try to do.

Thank Sergei for your quick response I am trying to output the following fields B,E,F,G by Month(C), B, E, F, G. The output looks like this:

B E F G
06962-5001375 INSURANCE IND ALL AC-EST -203.64
06962-5001375 INSURANCE THE PERSONAL -469.88
06962-5001375 OD HANDLING FEE 1 @ $ 5.00 -5
06962-5001375 Transfer WWW TRANSFER - 0982 1000.00

DATA
A B C D E F G
1 Chequing 06962-5001375 2023-08-01 " " INSURANCE IND ALL AC-EST -203.64
2 Chequing 06962-5001375 2023-08-01 " " INSURANCE THE PERSONAL -269.88
3 Chequing 06962-5001375 2023-08-01 " " INSURANCE THE PERSONAL -200.00
4 Chequing 06962-5001375 2023-08-02 " " OD HANDLING FEE 1 @ $ 5.00 -5
5 Chequing 06962-5001375 2023-08-09 " " Transfer WWW TRANSFER - 0982 1000.00
6 Chequing 06962-5001375 2023-08-09 " " Payment - 5925 MILTON - TAXE -50
7 Chequing 06962-5001375 2023-08-09 " " Payment - 2317 RELIANCE -55
8 Chequing 06962-5001375 2023-08-09 " " Payment ENT - 8635 MILTON HYDRO -145.52
Chequing 06962-5001367 2023-04-18 OD INTEREST -3.7

i removed www for readability only.

I hope I have got it right this time. I have been trying to get this right for at least two weeks and it is driving me crazy.

@Norman_McIlwaine 

Sorry, I still can't understand how the source data is structured. Is that as

Screenshot 2023-08-19 183227.png

? If so what is the logic behind generating these 4 output records, why these ones not another? Or your output doesn't cover entire input?

@Sergei Baklan 

This is the entire Source data, I need to have the data structured so that I can have B, F, G By G
i.e.
Output
06962-5001375 INSURANCE IND ALL AC-EST -203.64
06962-5001376 INSURANCE THE PERSONAL -469.88
06962-5001377 OD HANDLING FEE 1 @ $5 5
..
..
..
..
06962-5001367 OD HANDLING INTEREST 3.7


Sorry for the confusion my snip & sketch does not want to paste the snip. However, the above is what my output should look like.
Thanks for the support and patience in helping me resolve this issue.

Chequing 06962-5001375 2023-04-18 PETRO-CANADA IDP PURCHASE - 4189 -29.42
Chequing 06962-5001375 2023-04-18 C-IDP PURCHASE-3770 NOFRILLS JOHN'S -33.8
Chequing 06962-5001375 2023-04-18 OD INTEREST -0.92
Chequing 06962-5001375 2023-04-19 C-IDP PURCHASE-2812 LUNCHBOX CAFE -2.95
Chequing 06962-5001375 2023-04-19 C-IDP PURCHASE-9840 NOFRILLS JOHN'S -4.58
Chequing 06962-5001375 2023-04-19 C-IDP PURCHASE-9364 FOOD BASICS 88 -44.36
Sergei Sorry for all the confusion, I am trying to have the columns B, C ,E, F, G Displayed by month, by E, by F, by G.
the formula does this but with the by month and by B, E, F, G only if I change the UNIQUE and TOROW statements to the applicable TAKE variable. i.e
=LET(r,B1:G6,
m,4,
x,E1:E6,
a,TAKE(r,,1),
b,INDEX(r,,2),
c,INDEX(r,,3),
d,INDEX(r,,4),
e,INDEX(r,,5),
f,INDEX(r,,6),
v,VLOOKUP(x,x,1,FALSE),
u,SORT(UNIQUE(v)),
HSTACK(u,MMULT(--(TOROW(v)=u),(MONTH(b)=m)*f)))

OUTPUT
C-IDP PURCHASE-2812 -2.95
C-IDP PURCHASE-3770 -33.8
C-IDP PURCHASE-9364 -44.36
C-IDP PURCHASE-9840 -4.58
OD INTEREST -0.92
PETRO-CANADA -29.42

if I change the UNIQUE to "a" and TOROW to "a" my output changes to:
06962-5001375 -116.03

is this the right way to look at the data or can you suggest something else. I am open for any suggestions as I have spent far to many hours on this subject.

@Norman_McIlwaine 

To add screenshot here, copy it into clipboard, here click on Insert Picture button

image.png

after that click by mouse at any place in grey area

image.png

Ctrl+V, picture shall be loaded from the clipboard, when it's ready click Done at bottom right.

 

Back to the sample, I added headers for better understanding, in general you may ignore them.

image.png

I guess we need to sum amount for identical Id, Date, Transaction and Type. That could be done by

=LET(
    range, $B$2:$G$10,
    Id, CHOOSECOLS(range, 1),
    date, CHOOSECOLS(range, 2),
    transaction, CHOOSECOLS(range, 4),
    type, CHOOSECOLS(range, 5),
    amount, CHOOSECOLS(range, 6),
    groupIndexes, {1,2,4,5},
    group, UNIQUE( CHOOSECOLS(range, groupIndexes ) ),
    VSTACK({"Id","Date","Transaction","Type","Amount"},
    HSTACK(
        group,
        BYROW( group,
            LAMBDA(v, SUM(
                FILTER(amount,
                    (Id = INDEX(v,1,1) )*
                    (date = INDEX(v,1,2) )*
                    (transaction = INDEX(v,1,3) )*
                    (type = INDEX(v,1,4) )
                )
            ) )
        )
    ) )
)

If we would like to filter for the specific year and month as they are defined here

image.png

slightly modified formula will be

=LET(
    range, $B$2:$G$10,
    year, $J$14,
    monthNumber, $J$15,
    Id, CHOOSECOLS(range, 1),
    date, CHOOSECOLS(range, 2),
    transaction, CHOOSECOLS(range, 4),
    type, CHOOSECOLS(range, 5),
    amount, CHOOSECOLS(range, 6),
    monthRange, FILTER(range, EOMONTH(date,0) = EOMONTH( DATE(year,monthNumber,1), 0 )),
    groupIndexes, {1,2,4,5},
    group, UNIQUE( CHOOSECOLS(monthRange, groupIndexes ) ),
    VSTACK({"Id","Date","Transaction","Type","Amount"},
    HSTACK(
        group,
        BYROW( group,
            LAMBDA(v, SUM(
                FILTER(amount,
                    (Id = INDEX(v,1,1) )*
                    (date = INDEX(v,1,2) )*
                    (transaction = INDEX(v,1,3) )*
                    (type = INDEX(v,1,4) )
                )
            ) )
        )
    ) )
)

Your initial formula sum amounts for the unique Id (only one column) in a given month. It's can't be adjusted directly for the few unique columns.

Another way is to create PivotTable, something like

image.png

Please check in attached file if that close to what you'd like to receive.

Thank you Sergei for the most impressive work you have done for me. In particular I like the slightly modified formula. I never thought to have year & month columns but it is very functional. As you can see I am new to EXCEL and willing to learn about various formulas and exciting things you can do with this program. For now I will have to finish my personal finances. Your work has surly helped me a long way. Just one thing I have to ask you, will i be allowed to copy this code from you and run it without typing it all out in case I make a mistake. I can cut and past it from the document however how do I run it excel do I have to create a macro or use VBA which i absolutely no nothing about either?
Once again thank you for your kind help. It has been a long three weeks since I started this project.

@Norman_McIlwaine , thank you for the feedback.

Above is just formula which you may edit directly in formula bar. If you check the file from previous post it's here

image.png

but practically not visible if you don't expand formula bar. For that click on any empty place within it and Ctrl+Shift+U. Or drag bottom border down.

image.png

Editing here use Alt+Enter to insert new line.

In addition you may install Excel Labs add-in from the Store

image.png

within it it'll be Advanced Formula Environment to work with formulae and not only.

@Sergei Baklan 

 

Thank you Sergei for the valuable information, I very much appreciate it. I used the formula on the data unfortunately the amount was not summed i.e. The Personal in rows 4 and 5. I tried to review the formula  only changing the year and date variables to correspond with the input in col J. The rest I could follow in a basic form however , I could not see where the calculation was taking place to give the sum amount.  is it  the LAMBDA I see that is were you do all the sum. However, I cannot figure it out at this time.

Norman_McIlwaine_0-1692726826706.png

 

Norman_McIlwaine_3-1692727251039.png

 

 

Norman_McIlwaine_2-1692727115707.png

 

Thanks Sergei for all the help given.

@Norman_McIlwaine 

Formula works, correctly, perhaps I misunderstood the logic. It sums the values if only each of the value in these 4 columns are the same:

image.png

For them only INSURANCE in column C is the same value, thus they are not summarised. Which exactly fields shall be the same to summarize amounts?

@Sergei Baklan 

 

Sorry for the misunderstanding, in a previous image you had it right.

 

Norman_McIlwaine_0-1692730748493.png

In a previous capture you got it just right the C and D where the same so it added the amount.

 

@Norman_McIlwaine Sergei, If I change the date(day) to a 1 The function works great, Is there anyway we can add in a piece of code to perform a unique when the function is finished  I have tried several ways to include this into the function but have failed every time. I even wrote a separate piece of code which I got from the You Tube and adapted it to test my data prior to inserting however I get the same results frustration. 

 

Norman_McIlwaine_0-1692973004511.pngNorman_McIlwaine_1-1692973089777.png

 

Norman_McIlwaine_2-1692973450389.png

Norman_McIlwaine_3-1692973503253.png

Can you please help me Sergei?

@Norman_McIlwaine 

Could you please clarify what shall be unique, or, perhaps, you could manually generate desired result? I see now rows which have the same values in all columns, thus formula returns an array as it is.

@Sergei Baklan

 

Original Data

Norman_McIlwaine_6-1692986442282.png

Norman_McIlwaine_0-1692983182581.png

The above formula will only work if the fields C,D,F are of the same value

Changed Original date to 2023-08-01 from 2023-08-02 

Norman_McIlwaine_7-1692986670868.png

Norman_McIlwaine_4-1692986159095.png

The above formula works great,

If the date in column D5 are the same as column D4 2023-08-01

the Type field is then UNIQUE and the amount is summed as well. 

 

I need the Transaction and Type fields to be unique. In other words I am trying to not have all three fields the same just two. 

 

Thanks for all the help you have provided to me.

@Norman_McIlwaine 

Okay, thank you. If the dates in D4 and D5 are different, which date shall we show combining these two records, or we don't show dates at all in output?

@Sergei Baklan 

I removed the dates and found the formula worked well except I could not select by month. I don't need the dates at all in the output since we should will be selecting by year month.

Norman_McIlwaine_2-1692996337435.pngNorman_McIlwaine_3-1692996463875.png

 

 

The end result is good.

How can I select the data by Year and Month?

@Norman_McIlwaine 

To exclude dates what we shall do in formula is to remove/change related indexes

image.png

This part of the formula

monthRange, FILTER(range, EOMONTH(date,0) = EOMONTH( DATE(year,monthNumber,1), 0 )),

is filtering source range for the selected month, we keep it unchanged

Finally entire formula is

=LET(
    range, $B$2:$G$10,
    year, $J$14,
    monthNumber, $J$15,
    Id, CHOOSECOLS(range, 1),
    date, CHOOSECOLS(range, 2),
    transaction, CHOOSECOLS(range, 4),
    type, CHOOSECOLS(range, 5),
    amount, CHOOSECOLS(range, 6),
    monthRange, FILTER(range, EOMONTH(date,0) = EOMONTH( DATE(year,monthNumber,1), 0 )),
    groupIndexes, {1,4,5},
    group, UNIQUE( CHOOSECOLS(monthRange, groupIndexes ) ),
    VSTACK({"Id","Transaction","Type","Amount"},
    HSTACK(
        group,
        BYROW( group,
            LAMBDA(v, SUM(
                FILTER(amount,
                    (Id = INDEX(v,1,1) )*
                    (transaction = INDEX(v,1,2) )*
                    (type = INDEX(v,1,3) )
                )
            ) )
        )
    ) )
)

which returns

image.png

Please check in No Dates sheet attached.

@Sergei Baklan 

I found this to be very good on the small amount of data. When I expanded the data over several months (8,4) I found the amount is summed over all months while the output from Id, Transaction, and Type are only selecting by the month requested.

Norman_McIlwaine_0-1693084597064.png

Example PETRO

1 best response

Accepted Solutions
best response confirmed by Norman_McIlwaine (Brass Contributor)
Solution

@Norman_McIlwaine 

Thank you. Do I understand correctly that csv62719 is you source file. If to avoid Power Query we copy/paste data from it into the sheet in main file, here by formula create massaged range and based on it and month/year/monthid parameters create aggregated result. Workflow is correct?

View solution in original post