Find number in excel row

• 
  
  
  
  Excel Howtos
  

Extracting numbers from text in excel [Case study]

• 
  Last updated on June 19, 2012

65 Responses to “Extracting numbers from text in excel [Case study]”

1. Gregor Erbach says:

   I have learned one important lesson in my career: don’t use Excel as a word processor!
   Excel is very ill-suited to handling tasks like pattern matching in text (regular expressions).
   For example, one could use a Perl script (from the CMD command line) to extract numbers from the text, something like

    
   perl -n -e «m/([d,.]*)/; print «$1n» IN > OUT

    
   The point I am trying to make is, if text processing is a substantial part of your tasks, it is worth the effort learning a tool that is more suited to the task.

   • I agree with what your saying about suitability of tools to task,

     but seriously perl -n -e “m/([d,.]*)/; print “$1n” IN > OUT

     Doesn’t just roll off your tongue 

     and very few PC’s would even have Perl installed

   • Interesting point. I would have used VBA or some variation of that to clean data. But as Hui says Perl is not something many (including me) have on our computers. Plus I suck at regular expressions. I have tried long and hard to learn them, but I guess my mind is not wired to understand how they work 🙁

   • Jim says:

     The VBA engine for regular expressions works fine in Excel and Access. I agree if you are doing a substantial amount of text cleaning then doing so exclusively in Excel is probably not the best course, but using regular expressions in Perl or any other flavor doesn’t really provide any advantage over using regex in Excel. At the end of the day, regex is regex (with the exception of minor syntactic differences).

     Chandoo, a lot of people are intimidated by regex just like people are intimidated by formulas or macros. Once you learn a handful of rules and concepts, you can really do some damage with regex, and not just for data cleaning. You can validate input, parse,  transform (e.g. firstname lastname —> lastname, firstname), etc. Once you pick it up you’ll see many opportunities where it’s useful and sometimes superior to another approach (and sometimes not). My brain is not wired to understand multiple nested formulas or complicated array formulas so I guess regex is just easier to me. I would add that the regular expression approach not only cleans this data, but is flexible enough to account for other variations…try it on the other records you have and see how well it performs. Either way, it’s great to know there are multiple approaches to solving the problem!

2. Bhavani Seetal Lal says:

   How to sign inn into chandoo.org???

3. Vijay Kumar says:

   Dear Chandoo…
   Awesome, you are such a genius, how can you think like «what the excel will think»? Such ideas implementation…Very great . But lastly, i just wanted to know that in your formula =SUMPRODUCT(COUNTIF(lstDigits,MID(B4,ROW($A$1:$A$200),1))), why we can’t use 1 on behalf of ROW($A$1:$A$200), obviously if we will input this formula somewhere in excel, it results 1. So what is the logic to use row function instead of simple 1.
   Please explain, i am eager to learn the logic of inputting or building a formulas. Can you suggest me the source like web url or books to build better formulas in our excel work
    

   • Luke M says:

     Within SUMPRODUCT, that bit of formula is actually producing an array of numbers from 1 to 200, eg
     {1,2,3,4,…200}
     To see this in the formula, highlight that portion within the SUMPRODUCT and hit F9. 

4. In the first example above, I would store lstNumbers as a Named Formula:

   lstNumbers ={0;1;2;3;4;5;6;7;8;9}

5. I do not know why you would not want to use a UDF — Here is the code for one I wrote that extracts Numbers, Letters, Commas and spaces to clean up data  — amend as you wish for ANY ASCII characters
   Function NumbersAndLettersOnly(instring) As String
   Dim StringLength As Integer ‘to hold string length
   Dim i As Integer ‘counter for loop
   Dim AsciiVal ‘to hold working character
   Dim WorkingString As String ‘to build output string
   instring = Trim(instring) ‘Drop leading & trailing spaces
   StringLength = Len(instring) ‘Count number of characters in string
   For i = 1 To StringLength ‘Loop thru each character in the string
   AsciiVal = Asc(Mid(instring, i, 1))
   If AsciiVal >= 48 And AsciiVal <= 57 Then ‘Numbers 0-9
   WorkingString = WorkingString & Chr(AsciiVal)
   ElseIf AsciiVal >= 65 And AsciiVal <= 90 Then ‘A-Z
   WorkingString = WorkingString & Chr(AsciiVal)
   ElseIf AsciiVal >= 97 And AsciiVal <= 122 Then ‘a-z
   WorkingString = WorkingString & Chr(AsciiVal)
   ElseIf AsciiVal = 46 Then ‘.
   WorkingString = WorkingString & Chr(AsciiVal)
   ElseIf AsciiVal = 32 Then ‘{space}
   WorkingString = WorkingString & Chr(AsciiVal)
   End If
   Next i
   NumbersAndLettersOnly = WorkingString ‘Return output to function
   End Function

   • @Cliff,

     In case you might be interested, here is a shorter UDF which does the same thing as the one you posted…

     [code]
     Function AlphaNumerics(ByVal InString) As String
       Dim X As Integer
       For X = 1 To Len(InString)
         If Mid(InString, X, 1) Like «[!A-Za-z0-9. ]» Then Mid(InString, X, 1) = Chr$(1)
       Next
       AlphaNumerics = Replace(InString, Chr$(1), «»)
     End Function
     [/code]

     • Kevin says:

       I used this one, because it was the smallest but most readable (to me).

       The quotes didn’t translate correctly into VBA when I cut and pasted, so I just deleted them and typed them in again. Also, this function does the opposite: it finds the search characters in the 1st set of quotes, and then substitutes a null character. In my case, I wanted the opposite, so I just added a NOT in front of Mid. Thanks Rick. 🙂

6. Hi Purna,

    
   I had a similar request earlier this week. Here’s the ‘quick & dirty’ solution I used. Keep in mind that, if there are a lot of different characters and text you need to remove from your text, this can get a little cumbersome. But  for removing  a few different text strings and characters from the same cell, this is pretty simple to understand.

   Using your example, to remove the space character , INR, and Rs, I would use this formula…
    

   =SUBSTITUTE(SUBSTITUTE(SUBSTITUTE(B4,» «,»»),»INR»,»»),»Rs»,»»)
    

   You start with =SUBSTITUTE(B4,» «,»») and then wrap it in another SUBSTITUTE  for each text string you want to remove.

7. Elias says:

   One more option that works withh all Excel versions.

   =1*
   MID(A1,
   MIN(FIND({0,1,2,3,4,5,6,7,8,9},A1&»0123456789″)),
   FIND(» «,A1&» «,MIN(FIND({0,1,2,3,4,5,6,7,8,9},A1&»0123456789»)))
   )

   Confirm with Cntrl+Shift+Enter

   Regards

8. Elias says:

   Other more with different approach.

   =1*
   MID(A1,
   MATCH(1,1*ISNUMBER(1*MID(A1,ROW($1:$255),1)),0),
   MATCH(10,1*MID(A1,ROW($1:$255),1))
   )

   Ctrl+Shift+Enter

   Regards

   • Isaac says:

     You just saved me from giving up haha. I tried many formulas that didn’t work, this one did it!!! Thank You very much

     • @Isaac

       Never give up
       The Chandoo.org Forums are a great source of asking «How do I …» questions and getting a result real quick
       http://chandoo.org/forums/

9. Okay, one more method. The following was posted some time ago by Lars-Åke Aspelin in some forum or newsgroup…
   =MID(SUMPRODUCT(—MID(«01″&A1,SMALL((ROW($1:$300)-1)*ISNUMBER(-MID(«01″&A1,ROW($1:$300),1)),ROW($1:$300))+1,1),10^(300-ROW($1:$300))),2,300)
   This is an array formula and has to be confirmed with CTRL+SHIFT+ENTER rather than just ENTER.
   It has the following (known) limitations:
   — The input string in cell A1 must be shorter than 300 characters
   — There must be at most 14 digits in the input string.
      (Following digits will be shown as zeroes.)
   Maybe of no practical use, but it will also handle the following two cases correctly:
   — a «0» as the first digit in the input will be shown correctly in the output
   — an input without any digits at all will give the empty string as output (rather than 0).

10. Jim says:

    I would probably use regular expressions in this case. The functions are:

    ************************************************

    Function RegExReplace(ReplaceIn, _
            ReplaceWhat As String, ReplaceWith As String, Optional IgnoreCase As Boolean = False)
        Dim RE As Object
        Set RE = CreateObject(«vbscript.regexp»)
        RE.IgnoreCase = IgnoreCase
        RE.Pattern = ReplaceWhat
        RE.Global = True
        RegExReplace = RE.Replace(ReplaceIn, ReplaceWith)
    End Function

    Function RegExFind(FindIn, FindWhat As String, _
            Optional IgnoreCase As Boolean = False)
        Dim i As Long
        Dim matchCount As Integer
        Dim RE As Object, allMatches As Object, aMatch As Object
        Set RE = CreateObject(«vbscript.regexp»)
        RE.Pattern = FindWhat
        RE.IgnoreCase = IgnoreCase
        RE.Global = True
        Set allMatches = RE.Execute(FindIn)
        matchCount = allMatches.Count
        If matchCount >= 1 Then
            ReDim rslt(0 To allMatches.Count — 1)
            For i = 0 To allMatches.Count — 1
                rslt(i) = allMatches(i).Value
            Next i
            RegExFind = rslt
        Else
            RegExFind = «»
        End If
    End Function

    ************************************************

    and the usage is:

    =RegExFind(B4,»d{2,7}(.)?d{2,7}(.d{2}|,0{1,2})?»)

    The expression d{2,7}(.)?d{2,7}(.d{2}|,0{1,2})? is a little twisted, but these data are pretty dirty, too! Just using this expression: d+ would match all but two of the sample records.

    Cheers,
    -Jim

11. Godsbod says:

    The obvious solution would have been to make a more structured questionnaire, with a seperate value field for currency and amount…

    but the challenge now arisiing is almost a worthy by-product…

12. • Elias says:

      @ Juan,

      That formula doesn’t work in this case.

      Regards

13. Carl says:

    I think the bigger problem is the one created by separating the numbers from the letters.  The goal was to figure out what people are making in their excel jobs, and now we have numbers but no reference to the currency, so you can’t average these numbers to figure out how much the average excel-wise employee is making.  We’d have to go back and determine the currency and calculate the exchange rate.  I think @Godsbod had it right.  The more important lesson is to think through the questionnaire in order to get more manageable results, which is what we really want.

14. […] week we discussed how to extract numbers from text in Excel using formulas. In comments, quite a few people suggested that using VBA (Macros) to […]

15. r says:

    @Elias
    your formulas are very beautiful … But I do not understand why to use a matrix of constants!
    use ROW(1:10)-1 and 1234567890 … better!
    Also your formulas do not work in cases similar to this ABCD 1234 EFGH … You need to make small changes … Meanwhile I propose this is shorter and also works in these cases:
    =LEFT(REPLACE(A1,1,MIN(FIND(ROW(1:10)-1,A1&57321^2))-1,),FIND(» «,REPLACE(A1,1,MIN(FIND(ROW(1:10)-1,A1&57321^2))-1,)&» «)-1)
    ah … 57321 is a pandigital number 🙂
    regards
    r

    • Elias says:

      @r nice to see you around.
      Other than the number of characters I don’t see any advantages using the row option over the array of constants. However, using constants avoid the use of array enter and the formulas don’t get mess if you insert or delete rows.
      Regards

16. r says:

    @elias
    you’re right but I can not look at the formula
    is only an aesthetic factor … impossible to say «beautiful» by a matrix of constants
    regards
    r
     

17. r says:

    @elias
    i have changed your formula … so it work with ABCD 1234 EFGH:
    =—MID(LEFT(A1,FIND(» «,A1&» «,MIN(FIND(ROW(1:10)-1,A1&57321^2)))-1),MIN(FIND(ROW(1:10)-1,A1&57321^2)),20)
    what do you think?
    regards
    r
     

    • Elias says:

      @r

      I like that one, but it doesn’t work with cases like this ABCD12345EFGH. So, what about this one?

      =—LOOKUP(9.9E+307,—LEFT(MID(A1,MIN(FIND(ROW(1:10)-1,A1&57321^2)),255),ROW(1:255)))

      Regards

      • Jim says:

        My RegEx solution above handles both ABCD 1234 EFGH and ABCD12345EFGH.

        Cheers,
        -Jim

18. r says:

    oh @elias … fantastic! i like it!
    and, why not so:
    =-LOOKUP(0,-LEFT(MID(A1,MIN(FIND(ROW(1:10)-1,A1&amp;57321^2)),255),ROW(1:255)))
    🙂
    r
     

19. r says:

    ufff … so:
    =-LOOKUP(0,-LEFT(MID(A1,MIN(FIND(ROW(1:10)-1,A1&57321^2)),255),ROW(1:255)))

    • Elias says:

      @r
      Yea!!! I like the use of negative sign in 2 different places.

      Regards

    • bashful says:

      Great! I was able to use this formula to extract the first number from HGVS (human genome variation society) nomenclature:

      c.247_248delCG 247

      p.Glu789_Arg790del 789

      The c. or p. are sometimes omitted and the numbers can be of variable length (1-12 digits).

      I would like to understand how this formula works. I basically know lookup, left, mid, right and find but this is over my head!

      Thank you!

20. […] Extract numbers from text using formulas […]

21. joel says:

    Mr. Chandoo,
    please ask you a help
    I want to have a code for Excel 2010, to list on a sheet all macros from a workbook excel
    please, help me in this

    thank you very much

22. Raj says:

    I want to separate in each column like 25 PF 105 E01 XXXX , from below examples , please help or write srinivasmr@rediffmail.com

    25PF105E01XXXX

    100SP105E02XXXX

23. Jim says:

    Hi Raj,

    Use the RegExReplace function I posted above and then use this:

    =RegExReplace(A1,»(^d{2,3})(w{2})(d{3})(w{3}(w{4}))»,»$1 $2 $3 $5″)

    where the text you want to parse is in cell A1. 

    And the result is:

    25 PF 105 XXXX
    100 SP 105 XXXX

    It makes the following assumptions:

    — The first part is at least 2 but nore than 3 digits
    — The second part is 2 characters
    — The third part consists of three digits
    — The fourth part consists of 4 alpha-numeric characters

    Cheers,
    -Jim
            

    • rao says:

      Hi

      could you tell me how to learn regex tell me websites to learn for me

      my mail id: raoexcel@gmail.com

24. Rahim Zulfiqar Ali says:

    =IF(ISERROR(MID(B11,FIND(«/»,B11,1)-2,10)),»»,(MID(B11,FIND(«/»,B11,1)-2,10)))

25. Santosh says:

    Hi,

    I have an issue where I have a strings of number and letter that are separted by comma. (Q123,125-129,QA123,127-129).

    I would like extend the numbers that are inside the hyphen so that it becomes
    (Q123,125,126,127,128,129,QA123,127,128,129)

    Can this be completed using just the formulas.

26. Vijay Verma says:

    Check this out…This seems to be the right thing to do..
    http://www.excelforum.com/excel-formulas-and-functions/653043-extracting-numbers-from-alphanumeric-strings.html
    This says following —
    Try this

    A1: abc123def456ghi789

    First, create a Named Formula
    Names in Workbook: Seq
    Refers to: =ROW(INDEX($1:$65536,1,1):INDEX($1:$65536,255,1))

    This ARRAY FORMULA removes ALL non-numerics from a string
    B1: =SUM(IF(ISNUMBER(1/(MID(A1,seq,1)+1)),MID(A1,seq,1)*10^MMULT(-(seq<TRANSPOSE(seq)),-ISNUMBER(1/(MID(A1,seq,1)+1)))))

    In the example, the formula returns: 123456789
     

27. antonio says:

    please, help
    I need code for Excel 2010 for list all the macros of book excel, in one sheet. name of the macro and sheet of the macro
    many thanks

    • @Antonio

      Try the following:

      Sub ListMacros()
' Code thanx to Bob Philips
Const vbext_pk_Proc = 0
Dim VBComp As Object
Dim VBCodeMod As Object
Dim oListsheet As Object
Dim StartLine As Long
Dim ProcName As String
Dim iCount As Integer

      Application.ScreenUpdating = False
      On Error Resume Next
      Set oListsheet = ActiveWorkbook.Worksheets.Add
      iCount = 1
      oListsheet.Range("A1").Value = "Macro"

      For Each VBComp In ThisWorkbook.VBProject.VBComponents
      Set VBCodeMod = ThisWorkbook.VBProject.VBComponents(VBComp.Name).CodeModule
      With VBCodeMod
      StartLine = .CountOfDeclarationLines + 1
      Do Until StartLine >= .CountOfLines
      oListsheet.[a1].Offset(iCount, 0).Value = .ProcOfLine(StartLine, vbext_pk_Proc)
      iCount = iCount + 1
      StartLine = StartLine + .ProcCountLines(.ProcOfLine(StartLine, vbext_pk_Proc), vbext_pk_Proc)
      Loop
      End With
      Set VBCodeMod = Nothing
      Next VBComp

      Application.ScreenUpdating = True
End Sub


      or have a look at: http://msdn.microsoft.com/en-us/library/office/dd890502%28v=office.11%29.aspx

      • antonio says:

        Many thanks Huis
        Excel 2010 give errors in the all the lines

        For Each VBComp In ThisWorkbook.VBProject.VBComponents
        Set VBCodeMod = ThisWorkbook.VBProject.VBComponents(VBComp.Name).CodeModule
        With VBCodeMod
        StartLine = .CountOfDeclarationLines + 1
        Do Until StartLine >= .CountOfLines
        oListsheet.[a1].Offset(iCount, 0).Value = .ProcOfLine(StartLine, vbext_pk_Proc)
        iCount = iCount + 1
        StartLine = StartLine + .ProcCountLines(.ProcOfLine(StartLine, vbext_pk_Proc), vbext_pk_Proc)
        Loop
        End With
        Set VBCodeMod = Nothing
        Next VBComp

        greetings

        • • antonio says:

            Many thanks
            I have tried but it did not give result for Excel 2010
            This codes are for Excel 2007 or for VB, but it do not work in Excel 2010

            Regards
            Antonio

29. Beth says:

    With the data I had, I was able to use your first approach with great success. In my case I did not have any commas or decimals to worry about; I was simply trying to extract a simple account number from a string of letter and numbers. So, after creating the named range for lstNumbers, I was able to use the starting position formula, the number of characters formula, and then used a MID function to get the values I needed.

    Thank you, thank you, thank you! 🙂

30. […] The next example is from Chandoo […]

31. Jimmy Luong says:

    This is great. Thanks a lot

32. I encountered exactly the same situation recently, here’s my approach to solve it.
    =MID(A2,MIN(FIND({0,1,2,3,4,5,6,7,8,9},A2&»0123456789″)),COUNT(0+MID(A2,ROW(INDIRECT(«1:»&LEN(A2))),1)))+0
    Ctrl Shift Enter

    Not working for all cases though, e.g. with more than 1 set of numbers

    Here’s my post for sharing:
    http://wmfexcel.wordpress.com/2014/09/06/how-excel-formula-can-save-your-time/

33. Umer says:

    i am using formula
    {=MIN(IFERROR(FIND(lstNumbers,A1),””))}
    to extract 61.49 from 61.49 mg but it is giving
    #value!

    please guide what is wrong

    • Thomas says:

      Please help me with this problem
      I have a data as below
      IMR 123 , IMR456 , IMR #378 & IMR @890, MQC 123, QC234
      Now there are many records as above row
      I want only Number after IMR word
      Example for above i need output as
      123
      456
      378
      890
      All in different rows

34. jpcpa says:

    Hi Sir. I just on how can I sn separate the name and TAx information number in a two columns given that data below.

    «SAVER’S DIGITAL HUB APPLINCE DEPOT (1 Condura ) VAT 215-003-620
    «»DIGITEL MOBILE PHIL. VMH SUN-0922-877-7139 VAT 215-398-626
    «»DIGITEL MOBILE PHIL. VMH SUN-0922-877-7139 VAT 215-398-626
    «»DAPO REST AND BAR VAT 217-078-338
    «»WILCON BUILDER’S DEPOT INC. (1 pc. Vinyl Adhesive WaterBase 1 Gallon &
    110 pc. Kent Vinyl WP Deluxe 6″»X36″» Natural Oak-YHT) VAT 221-252-819
    «»WILCON BUILDERS DEPOT INC. VAT 221-252-819
    «»AVEXSON CORPORATION VAT 221-764-997 «
    «AVEXSON CORPORATION (for ZDJ 514) VAT 221-764-997
    «»TOKYO TERIYAKI RESTAURANT INC. VAT 226-172-350
    «»TRG AUTO GEM PARTS SUPPLY NV234-220-733»

    Thanks Sir.

    • @JCPCA

      What was wrong with my previous answer?
      Please ask the question in the Forums and attach a sample file
      http://chandoo.org/forum/

35. Martin says:

    Hi you commented that:
    As you can see, this method works well, but fails in some cases like,
    European number formats (, for decimal point and . for thousands)
    Text with multiple numbers

    Fortunately, in my data set, we had only a few incidents like these. So I have decided to manually adjust them than work out even more complicated formula.

    This is my problem, almost all my data is in this format and I need a formula to sort out this very issue…

36. kimdom says:

    I just cant get myself to think excel logic. Somebody help reading the number of days from a list like:

    Nov15 PUTS (23 days)
    March15 TIKS (3 days)
    March1 TIKS (25 days)
    June11 TIKS (10 days)

    So the reader should return the following:
    23
    3
    25
    10

    etc.

    • Hi Kimdom…

      Try this (assuming the text value in A1)

      =SUBSTITUTE(MID(A1, FIND(«(«,A1)+1,99),» days)»,»»)+0

      • kimdom says:

        Oh yes! Great!

37. Shlomi says:

    Good Day all,

    could not find a solution’s on all above comment….maybe someone can help me with this ?

    i have in a cell string such as 1.2.3.4 and i want to find if this string contains a number appear in other cell, lets say cell $A$1 = 3.
    I’ve able to do it, BUT
    when i have string such as: 1.2.13.4 , it looks on the 3 (on $A$1) and returned as if this number is part of the string.
    But i want to check if it contain 13 NOT 3.

    I being thinking that maybe I should check first the Length on the A1 cell contain the number (3) and then check the cell with the string (1.2.13.4).
    could not figure out how to do it….

    Try formula SEARCH or even by VBA , Instr(A1,»1.2.13.4″), but with no success. 🙁

    any help will be appreciated!!!

    Shlomi

38. TMR says:

    Thank you very much for your wonderful lessons for excel beginners like me.
    I have a question, if you have time, can you take a look and help me.
    Everyday I have to download a list of products which are sold in my site.
    Unfortunately, the site keep adding number to my product number to make it unique value to the system.
    For example, my product number is A1.
    In the first day, if 2 items of A1 were sold, then the system will shows A1-1, A1-2.
    And sometimes, if I mistakenly use the description of the listing for A1-1 as a sample to list new listing then the model number will be come A1-1-2

    So I will have a list of A1, A1-1, A1-2, A1-1-2, A1-1-2-1… for example.
    I want to delete all the added parts after A1.
    Can you help me with this problem?

    Thank you very much for your time

    • @TMR
      I’ll assume the A1-?? are in Column A
      Use a helper Column =Left(A1,2) then copy that down
      If The A1 can be A11 etc use =Left(A1,Find(«-«,A1)-1)
      then copy that down

      • TMR says:

        Thank you Hui for your reply.
        I see your answer and it really works with Product which has only 2 character like A1.
        But I also have some other products which has number: FA-3MF-BK
        or GFT-HJKU-25-BKJ
        So when I download the file from the system it always like this:
        FA-3MF-BK-1-2
        FA-3MF-BK-3
        GFT-HJKU-25-BKJ-1-1-2
        GFT-HJKU-25-BKJ-2-1

        In this case, what should I do?

        • @TMR
          How about posting the question in the Chandoo.org Forums
          https://chandoo.org/forum/

          Please attach a sample file with example of the range of input and outputs you expect?

          Otherwise we will be going around in circles for weeks

39. Jeff Faris says:

    I need to be able to extract a date from this formatted lot number. F19021286Z. I don’t need the F or the Z. 19 (year) 02(month) 12(day) 86 (last two digits are sequence numbers and are un needed.

    Do you have a formula that can do all of that?

Leave a Reply