I have the dynamic range A (A#) and I want to calculate the cumulative sum of each "subrange" in B. The result musst be something as shown in column B. In other words I need a dynamic function/ formula which recognises the consecutive numbers in A und computes their total each time.
I only managed to solve this with a helper column C
C1: 1
C2: =IF(A1=A2,C1,C1+1)
C2 is draggable all the way down where you need it
B1: =IF(A1=A2,"",SUMIFS(A1,A1,A1))
B2: =IF(A2=A3,"",SUMIFS($A$1:A2,$C$1:C2,C2,$A$1:A2,A2))
B2 is draggable all the way down where you need it
With Office 365 you can use this LET to create a dynamic array that spills into Column B as you have shown:
=LET( a, A1#,
sa, SCAN( 0, a, LAMBDA(a,b, a + b ) ),
rSeq, SEQUENCE( ROWS( a ) ),
a_1, INDEX( a, rSeq + 1 ),
change, IFERROR( a_1 <> a, 1 ),
sChange, SCAN( 0, change*sa, LAMBDA(a,b, MAX( a, b ) ) ),
IF( change, sChange - IF( rSeq=1, 0, INDEX( sChange, rSeq - 1) ), "" ) )
where A1# is the input array.
Can someone please help me with an excel formula or macro to extract the date and time in the 2nd last row? This text is part or a ticket dump export and I need date and time stamp of when it was first acknowledged. The names can vary and so can the lines of text, however, I would always need the date and time in the 2nd last line. Please help!
2021-08-01 19:17:21 - Kamatham Rakesh (Additional comments)
sample text
2021-08-01 19:04:48 - Shelley Black (Additional comments)
Sample Text
2021-08-01 19:04:31 - Kamatham Rakesh (Additional comments)
Sample text
2021-08-01 19:04:30 - Shelley Black (Additional comments)
Sample text
2021-08-01 19:04:05 - Shelley Black (Additional comments)
Sample text
2021-08-01 19:03:41 - Shelley Black (Additional comments)
Sample text
2021-08-01 19:03:25 - Shelley Black (Additional comments)
Hello
2021-08-01 19:03:17 - Kamatham Rakesh (Additional comments)
Hi shelly
2021-08-01 19:03:03 - Kamatham Rakesh (Additional comments)
Thank you for waiting. Sample text
2021-08-01 19:02:29 - Shelley Black (Additional comments)
sample text
OK - if all text is in a single column, this may work:
=INDEX( LEFT( FILTERXML( "<main><sub>" & SUBSTITUTE( A2, CHAR(10), "</sub><sub>" ) & "</sub></main>", "//sub" ), 19 ),
ROWS( FILTERXML( "<main><sub>" & SUBSTITUTE( A2, CHAR(10), "</sub><sub>" ) & "</sub></main>", "//sub" ) )-2 )
This assumes that the lines are separated by 2 carriage returns (i.e., CHAR(10)). It is delivering the text of the datetime.
If you want to return the datetime value, it may be a lot easier to deploy this with a helper cell, but you can also do this:
=DATEVALUE( INDEX( LEFT( FILTERXML( "<main><sub>" & SUBSTITUTE( A2, CHAR(10), "</sub><sub>" ) & "</sub></main>", "//sub" ), 19 ),
ROWS( FILTERXML( "<main><sub>" & SUBSTITUTE( A2, CHAR(10), "</sub><sub>" ) & "</sub></main>", "//sub" ) )-2 ) )
+ TIMEVALUE( INDEX( LEFT( FILTERXML( "<main><sub>" & SUBSTITUTE( A2, CHAR(10), "</sub><sub>" ) & "</sub></main>", "//sub" ), 19 ),
ROWS( FILTERXML( "<main><sub>" & SUBSTITUTE( A2, CHAR(10), "</sub><sub>" ) & "</sub></main>", "//sub" ) )-2 ) )
One way to do it would be:
=DATEVALUE( LEFT( INDEX( A:A, LARGE(IF( ISTEXT(A:A), ROW(A:A) ),2) ), 19) )
+ TIMEVALUE( LEFT( INDEX( A:A, LARGE(IF( ISTEXT(A:A), ROW(A:A) ),2) ), 19) )
This assumes the data are laid out in Column A of some arbitrary length:
This might take a lot of time to calc if your data are really long. There could be more calc-efficient ways if you knew exactly how many rows are in the data set. Unfortunately, the IF has to be repeated because the TIMEVALUE and DATEVALUE are computed separately and in Excel 2016, you don't have access to LET or FILTER which would speed things up a bit.
I have a table filled with cells formatted like this:
---(cell)---
3 x item
2 x another item
299 x yet another item
(... - variable amount of items here)
4 x the last item in this cell
---(end of cell)---
All items are separated with carriage returns (=CHAR(10)).
I need to extract the numbers and do various operations on them according to the item type. I'll take it one problem at a time and try to extract the numbers first. I'd use MID but it doesn't help with finding "the first set of numerical characters after a carriage return".
Any idea?
Here's a collection of formulas that will get you started. I have the cell to analyze as A1, and table of analysis results in B7:I11. The columns B-I are Item#, Start, End, SubString, Delim, Quantity and Item.
First, you can count the number of s in your cell
nCRs: =LEN($A$1)-LEN(SUBSTITUTE($A$1,CHAR(10),""))
The number of items is the number of CRs plus 1:
nItems: =LEN($A$1)-LEN(SUBSTITUTE($A$1,CHAR(10),""))+1
Note: to use the same tools on all items, I'm adding a CR to the beginning and end of the string. That way I'm always extracting sub-strings that are CR-delimited. To get the first item, put a 1 in B7:
B7: 1
To get the start of the first sub-string, use SEARCH with start pos = 1
C7: =SEARCH(CHAR(10),CHAR(10)&$A$1&CHAR(10),1)
Similarly get the end of the first sub-string with SEARCH starting to the right of where the first CR was found
D7: =SEARCH(CHAR(10),CHAR(10)&$A$1&CHAR(10),C7+1)
Now use MID to extract the text between two
E7: =MID(CHAR(10)&$A$1&CHAR(10),C7+1,D7-C7)
In F7 I have delimiter text (between item count and item description), assumed to be " x ":
F7: =" x "
In G7 find the location of the delimiter
G7: =SEARCH(F7,E7,1)
In H7 extract the text left of the delimiter (item count) and convert to a number
H7: =VALUE(LEFT(E7,G7-1))
In I7 extract the item description
I7: =RIGHT(E7,LEN(E7)-G7-LEN(F7)+1)
To get the second (and subsequent) items, start the search for CR where the last CR was found. So in C8
C8: =SEARCH(CHAR(10),CHAR(10)&$A$1&CHAR(10),D7)
Fill the formulas down to complete the table.
Hope that helps
This solution assumes data with text string containing multiple items per cell and separated by carriage returns is located in Sheet1 at B6:C11 as shown in figure 1 (adjust ranges in formulas as required).
Fig. 1
Objective as presented by OP:
I need to extract the numbers and do various operations on them
according to the item type. I'll take it one problem at a time and try
to extract the numbers first.
Yes, they're (numbers) always at the start of each carriage return, as
exemplified. What I need to do is stuff like "sum all numbers", "sum
all item A", "sum all item B" etc.
This solution proposes a review the objectives in order to first work on the extraction of the items to create a table with all items, after this is achieved then all other goals can be workout in a simpler manner.
First we add a column to the data in Sheet1. Enter this formula in C7 and copy till last record.
=SUM(1+LEN($B7),-LEN(SUBSTITUTE($B7,CHAR(10),"")))
Let’s create in Sheet2 a table to extract the items from the data in the Sheet1.
The extraction table is located at B6:G55 and contains the following items (see Fig. 2):
Itm.Nbr(+) : Enter this formula in B7 and copy till last record
=IF( OR( B6 = SUM( Sheet1!$C$6:$C$11 ), B6 = "" ), "",
- 1 + ROWS( B$6:B7 ) )
Line(+) : Enter this formula in B7 and copy till last record
=IF( EXACT( C6, C$6 ), 2,
IF( $B7 = "", "",
IF( $B7 <= SUM( INDEX( Sheet1!$C$6:$C$11, 1 ) : INDEX( Sheet1!$C$6:$C$11, C6 ) ), C6,
SUM( 1, C6 ) ) ) )
Itm.Lne(+) : Enter this formula in B7 and copy till last record
= IF( EXACT( D$6, D6 ), 1,
IF( $B7 = "", "",
IF( $C7 = $C6, SUM( 1, D6 ), 1 ) ) )
Extracted Lines(+) : Enter this formula in B7 and copy till last record
=IF( $B7 = "", "",
CLEAN( IF( $D7 = 1,
LEFT( INDEX( Sheet1!$B$6:$B$11, $C7 ),
FIND( CHAR(10), INDEX( Sheet1!$B$6:$B$11, $C7 ) ) ),
MID( LEFT( INDEX( Sheet1!$B$6:$B$11, $C7 ) & CHAR(10),
SEARCH( CHAR(135), SUBSTITUTE( INDEX( Sheet1!$B$6:$B$11, $C7 ) & CHAR(10), CHAR(10), CHAR(135), $D7 ) ) ),
SEARCH( CHAR(135), SUBSTITUTE( INDEX( Sheet1!$B$6:$B$11, $C7 ), CHAR(10), CHAR(135), - 1 + $D7 ) ), LEN( INDEX( Sheet1!$B$6:$B$11, $C7 ) ) ) ) ) )
Item : Enter this formula in B7 and copy till last record
=IF( $B7 = "", "",
IFERROR( MID( $E7, 3 + SEARCH( " x ", $E7 ), LEN( $E7 ) ), "!Err" ) )
Qty : Enter this formula in B7 and copy till last record
=IF( $B7 = "", "",
IFERROR( --TRIM( LEFT( $E7, FIND( " ", $E7 ) ) ), "!Err" ) )
(+) This working fields can be hidden
Fig. 2
I have an excel file that has been uploaded here
http://www58.zippyshare.com/v/99974349/file.html
The formula works great except for a column with descending values.
=INDEX(
INDIRECT("'"&LOOKUP(B5,TblA)&"'!A6:A36"),
LOOKUP(9.99999999999999E+307,
SEARCH("-"&C8&"-","-"&INDIRECT("'"&LOOKUP(B5,TblA)&"'!C6:C36")&"-"),
ROW(C6:C36)-ROW(C6)+1))
Let me explain the excel file.
I have one main sheet 'Report' and 4 other sheets correspond to 4 age groups. - 4.2.0 to 4.7.30, 4.8.0 to 5.1.30, 5.2.0 to 5.7.30 and 5.8.0 to 6.1.30. Depending on the Age (B5) in the sheet 'Report', I select one of the 4 sheet to pick values from. I pick the correct sheet using a Table Name TblA which contains all sheet names and is defined from A24 to B27 in the sheet 'Report'.
In the sample sheet that is uploaded, B5 contains the value 5.7 which means we have to select the sheet 5.2.0 to 5.7.30.
Now from the sheet 5.2.0 to 5.7.30, I have to seek the respective Standard Score (1st column) for every Raw Score entered in 'Report'.
Here are the steps:
A. Enter Raw scores in sheet 'Report' C7 to C15
B. Search Respective sheet depending on age (B5 cell), in our case 5.2.0 to 5.7.30 since age is 5.7
C. Populate Standard score from Raw scores by picking the corresponding column in the 4 sheets. For example, if Raw Score of Col1 is 25 (C7), then pick the Standard score of Col1 from 5.2.0 to 5.7.30 and enter in D7 and so on.
D. This way all standard scores are filled in D7 to D15.
The formula works great except for D13 in sheet 'Report' since if you observe ColD in 5.2.0 to 5.7.30, it is in descending order.
How do I change the formula to accomodate this unique column?
Well, it's not really the order that's causing the error, it's because you don't have any results! The formula you use is trying to find -159- which it cannot find at all in the age sheet. You really need something to look into ranges, so that if you have 159, it will return a positive result when you try to match against 139-160.
I have made a formula building it from smaller ones, but when assembled, the repeating units make it daunting... Also, it's an array formula, so you need to use Ctrl+Shift+Enter for it to work as intended. You can still drag the formula down.
=INDEX(
INDIRECT("'"&LOOKUP($B$5,TblA)&"'!A6:A36"),
IFERROR(
MATCH(
C7,
INDEX(
INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),
0,
MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)
)*1,
0
),
MATCH(
1,
IF(
1*LEFT(
INDEX(
INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),
0,
MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)
),
FIND(
"-",
INDEX(
INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),
0,
MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)
)
)-1
)<=C7,
1,
0
)*
IF(
1*MID(
INDEX(
INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),
0,
MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)
),
FIND(
"-",
INDEX(
INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),
0,
MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)
)
)+1,
100
)>=C7,
1,
0
)
,0
)
)
)
The single line version...
=INDEX(INDIRECT("'"&LOOKUP($B$5,TblA)&"'!A6:A36"),IFERROR(MATCH(C7,INDEX(INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),0,MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0))*1,0),MATCH(1,IF(1*LEFT(INDEX(INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),0,MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)),FIND("-",INDEX(INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),0,MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)))-1)<=C7,1,0)*IF(1*MID(INDEX(INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),0,MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)),FIND("-",INDEX(INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),0,MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)))+1,100)>=C7,1,0),0)))
You can notice that there are some repeating blocks, namely:
INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36")
For the sheet name;
INDEX(
INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),
0,
MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)
)
Which is a larger block to make the formula a bit more flexible (it automatically picks the correct column e.g. if you change B8 Exclusion to Col1, the formula will automatically adjust itself)
If I call the first Sheet and the second Column, it becomes much shorter and perhaps easier to understand:
=INDEX(
Sheet,
IFERROR(
MATCH(
C7,
Column*1,
0
),
MATCH(
1,
IF(
1*LEFT(
Column,
FIND(
"-",
Column
)-1
)<=C7,
1,
0
)*
IF(
1*MID(
Column,
FIND(
"-",
Column
)+1,
100
)>=C7,
1,
0
)
,0
)
)
)
Or
=INDEX(Sheet,IFERROR(MATCH(C7,Column*1,0),MATCH(1,IF(1*LEFT(Column,FIND("-",Column)-1)<=C7,1,0)*IF(1*MID(Column,FIND("-",Column)+1,100)>=C7,1,0),0)))
Disclaimer: I'm not sure if there are any way to make this even shorter, but I guess that as long as it's working right now ^^
You can download your updated sheet here.
Explanation:
As I mentioned before, the formula is based off several smaller ones and quite a few repeats of those.
INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36")
As you already know (it's a variation of a part of your own formula), this gives the area containing all the different ages. Using it and the below, we get this:
INDEX(
INDIRECT('"&LOOKUP($B$5,TblA)&"'!B6:J36"),
0,
MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)
)
Into:
INDEX(
'Sheet'!B6:J36,
0,
MATCH(B7,'Sheet'!B5:J5,0)
)
Index will thus look into the range 'Sheet'!B6:J36, 0 indicates it will take any column(s) and MATCH(B7,'Sheet'!B5:J5,0) returns the nth column by taking the value of B7 (in the case of your spreadsheet, Col1) and looking it into 'Sheet'!B5:J5 which gives 1. The above will thus return the range 'Sheet'!B6:B36. Let's put it in the formula:
=INDEX(
'Sheet'!A6:A36,
IFERROR(
MATCH(
C7,
'Sheet'!B6:B36*1,
0
),
MATCH(
1,
IF(
1*LEFT(
'Sheet'!B6:B36,
FIND(
"-",
'Sheet'!B6:B36
)-1
)<=C7,
1,
0
)*
IF(
1*MID(
'Sheet'!B6:B36,
FIND(
"-",
'Sheet'!B6:B36
)+1,
100
)>=C7,
1,
0
)
,0
)
)
)
This formula is itself a giant INDEX formula, with range 'Sheet'!A6:A36 and row number as the big IFERROR group. The first part of the IFERROR() gets evaluated first:
MATCH(
C7,
'Sheet'!B6:B36*1,
0
)
This should be easy enough to understand. It looks for the raw score (from C7) into the range we obtained earlier, times 1 to convert everything to number (you can't look up numbers and text and expect a match). So that if there's an exact match of a number, it will return the row number of the found raw score and feed it to the INDEX(). For example, if the first row is returned, we get:
=INDEX('Sheet'!A6:A36,1)
Which is 'Sheet'!B6. If however there's no match (i.e. the raw score cannot be found), MATCH will return an error. And that's when the second part of the IFERROR comes into play:
MATCH(
1,
IF(
1*LEFT(
'Sheet'!B6:B36,
FIND(
"-",
'Sheet'!B6:B36
)-1
)<=C7,
1,
0
)*
IF(
1*MID(
'Sheet'!B6:B36,
FIND(
"-",
'Sheet'!B6:B36
)+1,
100
)>=C7,
1,
0
)
,0
)
This MATCH tries to find 1 within what seems to be two IFs; the first one being:
IF( 1*LEFT('Sheet'!B6:B36,FIND("-",'Sheet'!B6:B36)-1)<=C7 , 1 , 0)
FIND("-",'Sheet'!B6:B36)-1 gets the position of the last character before the - in the column 'Sheet'!B6:B36.
With those values, this FIND would return:
12-13 -> 2
145-155 -> 3
1567-1865 -> 4
The IF thus becomes:
IF( 1*LEFT('Sheet'!B6:B36,{2,3,4})<=C7 , 1 , 0)
Notice the braces here; they indicate an array and that's why this is an array formula. LEFT then extracts all the characters before the - (remember your other question, I answered with a technique very similar to this):
12-13 -> 2 -> 12
145-155 -> 3 -> 145
1567-1865 -> 4 -> 1567
Which is...
IF( 1*{12,145,1567}<=C7 , 1 , 0)
Again, 1* converts those to actual numbers because LEFT be default returns text characters. It's important here to do this because we're going to use the comparator <=, so that if the value to the left of C7 (the raw score), then the IF should return 1, else, it should return 0. Let's say that the raw score was 154. The results would be:
IF( {12,145,1567}<=154 , 1 , 0)
IF( {TRUE,TRUE,FALSE} , 1 , 0)
{1,1,0}
I just realised that the formula can be made a little shorter xD Anyway, we'll see that later. The next IF behaves in a similar fashion, but checks for the value at the right of the -:
IF( 1*MID('Sheet'!B6:B36,FIND("-",'Sheet'!B6:B36)+1,100)>=C7 , 1 , 0)
With...
FIND MID('Sheet'!B6:B36, X, 100)
12-13 -> 4 -> 13
145-155 -> 5 -> 155
1567-1865 -> 6 -> 1865
You can notice that this formula will stop working if you have something above 100 character long here. Anyway, the IF thus becomes:
IF( {13,155,1865}>=154 , 1 , 0)
IF( {FALSE,TRUE,TRUE} , 1 , 0)
{0,1,1}
Now that we have these, the MATCH from before becomes:
MATCH( 1 , {1,1,0}*{0,1,1} , 0)
Some simple math makes this into:
MATCH( 1 , {0,1,0} , 0)
And what is the position of the 1 in there? That's right, position 2!
Our original formula this becomes:
=INDEX( 'Sheet'!A6:A36 , IFERROR( #Error! , 2 ) )
So that if nothing was found at first, it will return an error (#N/A in this case) and instead return 2. =INDEX( 'Sheet'!A6:A36 , 2 ) gives 'Sheet'!A7.
And the slightly shorter version is:
=INDEX(INDIRECT("'"&LOOKUP($B$5,TblA)&"'!A6:A36"),IFERROR(MATCH(C7,INDEX(INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),0,MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0))*1,0),MATCH(1,(1*LEFT(INDEX(INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),0,MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)),FIND("-",INDEX(INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),0,MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)))-1)<=C7)*(1*MID(INDEX(INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),0,MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)),FIND("-",INDEX(INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B6:J36"),0,MATCH(B7,INDIRECT("'"&LOOKUP($B$5,TblA)&"'!B5:J5"),0)))+1,100)>=C7),0)))
I actually removed the inner IFs, because (a>b)*(c>b) already returns 0s and 1s since TRUE multiplied by TRUE gives 1 in excel.
In Excel 2007 the formulae are returning a lot of circular reference warnings, so it may be worth adding a tag for your Excel version. Mine is Excel 2007 but with it the results you want seem achievable as below:
To shorten the formulae and reduce computation I have added in Report C5 "Table" and in D5 =VLOOKUP(B5,TblA,2,1).
I have also inserted a column immediately to the right of ColumnH ("ColD") in 5.2.0 to 5.7.30 and applied Text To Columns on Column H, with - as the delimiter.
I then applied to Report E7 and copied down to E15:
=INDEX(INDIRECT("'"&D$5&"'!A6:A36"),MATCH(C7,INDIRECT("'"&D$5&"'!"&CHAR(ROW()+59)&"6:"&CHAR(ROW()+59)&"36"),0))
and adjusted the 59s to 60s in the last three rows. Such adjustment would not be necessary if ColumnI were moved far enough to the right.
In E13 I changed the match from exact to next higher (final 0 to -1).
For Figures I cheated and changed K23 in 5.2.0 to 5.7.30 to 22 from 21-22, but such banding could, for other columns, be treated in much the same way as I did for ColD.