Getting Rid of Everything Except Numbers

Formula doesnt work

Or another version:

=TEXTJOIN("",TRUE,IFERROR(VALUE(MID(A1,SEQUENCE(LEN(A1)),1)),""))

(for them's what's got the SPILL functionss). Note that using "LEN(A1)" in the SEQUENCE() function makes it equal to the "ROW(INDIRECT..." approach in that exactly how many characters need checked will be checked, no more, no less.

And:

=TEXTJOIN("",TRUE,IFERROR(VALUE(MID(A1,ROW(INDIRECT("1:"&LEN(A1))),1)),""))

(for them's what's not got the SPILL functions)


Neither one though lets anything but the digits 0-9 through.

To get decimal points, one could try:

=TRIM( TEXTJOIN( ,FALSE, UNICHAR( IFERROR( IF( 1 * (UNICODE(MID($A$1,SEQUENCE(LEN($A$1)),1)) > 44) * (UNICODE(MID($A$1,SEQUENCE(LEN($A$1)),1)) < 58) * UNICODE(MID($A$1,SEQUENCE(LEN($A$1)),1)) = 0, 32, 1 * (UNICODE(MID($A$1,SEQUENCE(LEN($A$1)),1)) > 44) * (UNICODE(MID($A$1,SEQUENCE(LEN($A$1)),1)) < 58) * (UNICODE(MID($A$1,SEQUENCE(LEN($A$1)),1)))), 32))))

or without SEQUENCE():

=TRIM( TEXTJOIN( ,FALSE, UNICHAR( IFERROR( IF( 1 * (UNICODE(MID($A$1,ROW(INDIRECT("$1:$"&LEN($A$1))),1)) > 44) * (UNICODE(MID($A$1,ROW(INDIRECT("$1:$"&LEN($A$1))),1)) < 58) * UNICODE(MID($A$1,ROW(INDIRECT("$1:$"&LEN($A$1))),1)) = 0, 32, 1 * (UNICODE(MID($A$1,ROW(INDIRECT("$1:$"&LEN($A$1))),1)) > 44) * (UNICODE(MID($A$1,ROW(INDIRECT("$1:$"&LEN($A$1))),1)) < 58) * (UNICODE(MID($A$1,ROW(INDIRECT("$1:$"&LEN($A$1))),1)))), 32))))

which will keep what one might readily think are digits meant to represent discrete numbers within the string together putting a space between each grouping as well as keep decimal points and negative signs. So "KMJ883.4dhe-23" would yield "883.4 -23". This way, if those values ARE really meant to be actual numbers buried in the string, not simply digits that happen to follow one another, they can be separated out with standard techniques. Also, it TRIM()'s the result so obnoxiousness that way is gone. (My need was to separate out part dimensions from strings that were often bizarrely made, and often included foreign language characters that were of no value to me for their use. Hence the desire to keep related digits together to be numbers not characters and hence the UNICODE()/UNICHAR() angle. Did not need negative signs, and I don't keep them, but changing it to be "> 44" instead of > 45" wasn't hard...

Seems like one could take a different approach vis-a-vis what characters are kept. Mine was informed by the fact they are all in a sequence numerically, in character sets. But one could easily substitute a lookup approach ( XLOOKUP(), VLOOKUP() ) and have a table with one column, or two, first one to have the characters to keep, second to make those characters to notice, then look in the second column for the actual character to use in the output. Just a little re-jigging and the list of characters could easily be added to, etc.

Recently someone wanted to produce two strings, one with all the alpha characters removed (replaced by "double quote-space-double quote" actually), and vice versa. With just spaces instead:

Numerals only, with next-to-each-other numerals grouped as numbers in the result:

=TRIM(TEXTJOIN("",TRUE,IFERROR(VALUE(MID(A1,SEQUENCE(LEN(A1)),1))," ")))

and the text instead, kept in any side-by-side groupings:

=TRIM(TEXTJOIN("",TRUE,IF(ISERROR(VALUE(MID(A1,SEQUENCE(LEN(A1)),1))),MID(A1,SEQUENCE(LEN(A1)),1)," ")))

and one could modify either to have no spaces, easily, by replacing the final part at the end (just before the last three ")" that end the formulas) with just a pair of double quotes: "" .

The formula to strip non-numerics also sees decimal points as non-numeric, and splits on the decimal.

Apologies for the double post but I got my version fixed.

'As Variant' instead of 'As String' solved it.

Not sure how speedy it is but I'm good with regular expressions so it's very customizable.

Function StripChar(Txt As String) As Variant
With CreateObject("VBScript.RegExp")
.Global = True
.Pattern = "\D"
StripChar = Val(.Replace(Txt, " "))
End With
End Function

Hi, The 1st fn works well. But I'm having trouble with this. It's using regexp. I just what the fn to return a value and not a string. Thanks in advance.

Function StripChar(Txt As String) As String
With CreateObject("VBScript.RegExp")
.Global = True
.Pattern = "\D"
StripChar = .Replace(Txt, "")
End With
StripChar = Val(Txt)
End Function

@To whom it may concern,
To my opinion - the suggested formula can be at least 50% shorter.
=SUM(MID(0&A1,LARGE(INDEX(ISNUMBER(--MID(A1,ROW($1:$256),1))*ROW($1:$256),),ROW($1:$256))+1,1)*10^ROW($1:$256)/10)
----------------------------
Michael (Micky) Avidan
“Microsoft® Answers" - Wiki author & Forums Moderator
“Microsoft®” Excel MVP – Excel (2009-2018)
ISRAEL

Thank you very much. The formula works!