VBA replace a string EXCEL 2019 - excel

Question

I cannot extract the postal/zip code of a given address cell that comes like this :
"108, avenue du Grand Sud 37 170 CHAMBRAY les TOURS".
I have used :
=RECHERCHE(9^9;--("0"&STXT(A2;MIN(CHERCHE({0.1.2.3.4.5.6.7.8.9};A2&"0 123456789"));LIGNE($1:$100))))
Which sometimes works, sometimes not depending on the street number starting the address (here "108,").
The problem is the space of the pattern "37 170". I would like to remove the blank space in the pattern. Is there a regex way to search this pattern "## ###", and then to remove this poisonous blank space?
Thank you for your tricks.
I have tried this piece of code :
Function toto(r, Optional u = 0)
Application.Volatile
Dim i%, j%, adr$, cp$, loca$, x
x = Split(r)
For i = 0 To UBound(x)
If x(i) Like "#####" Then Exit For
Next
If i > UBound(x) Then
adr = r.Value 'facultatif
Else
cp = x(i)
For j = 0 To i - 1: adr = adr & x(j) & " ": Next
adr = Left$(adr, Len(adr) + (Len(adr) > 1))
For j = i + 1 To UBound(x): loca = loca & x(j) & " ": Next
loca = Left$(loca, Len(loca) + (Len(loca) > 1))
End If
x = Array(adr, cp, loca)
If 0 < u And u < 4 Then toto = x(u - 1) Else toto = x
End Function
The above code works fine for splitting addresses including street number, zip code, and city name. But it does not work when the zip code is ## ### = 2 digit integer - space - 3 digit integer.
Edit: 01 June 2021
Since it seems my question is not clear enough, let's rephrase :
Given an Excel worksheet containing in each cell of column A, from saying A1 down to A10000, complete addresses like this one :
"2 rue Rene cassin Centre Commercial Châlon 2 Sud 71 100 CHALON SUR SAONE"
or this one :
"15, Rue Emile Schwoerer 68 000 COLMAR"
Where "71 100" and "68 000" are a zip code in incorrect format because of the extra space between the 2 first digits and 3 last digits.
I need to split the Ai cell content in order to obtain :
in cell Bi : the text (street, etc.) placed left before the 2 first digits of the "wrong" zip code,
in cell Ci : the zip code with its correct format ("71100" and not "71 100"),
in cell Di : the text (city name) after the zip code.
It's a kind of left and right extraction around the zip code.
The above code that I have posted does not work.
In order to obtain the correct zip code format, I have tried the regex following function :
Function FindReplaceRegex(rng As Range, reg_exp As String, replace As String)
Set myRegExp = New RegExp
myRegExp.IgnoreCase = False
myRegExp.Global = True
myRegExp.Pattern = reg_exp
FindReplaceRegex = myRegExp.replace(rng.Value, replace)
End Function
But I am unable to determine the correct regular expression pattern to get rid of the space in the zip code.
PEH gave me the following pattern :
(.*)([0-9]{2} ?[0-9]{3})(.*)
When using the function, I have tried to define the replacement pattern by:
(.*)([0-9]{2}[0-9]{3})(.*)
But it would not work. Hope this will clarify my question.
Any idea is welcome. Thanks

Answer 1

If these input strings always have the same pattern, try:
=CONCAT(FILTERXML("<t><s>"&SUBSTITUTE(A1," ","</s><s>")&"</s></t>","//s[.*0=0]"))
Depending on your needs/edge-cases, you could add more xpath expressions.

Answer 2

If this is VBA, I have a fix for you (please forgive the crappy naming convention, I'm scribbling this down in work while waiting for SQL to refresh):
Sub test1()
a0 = Cells(1, 1) 'Get the text, in this case "108, avenue du Grand Sud 37 170 CHAMBRAY les TOURS"
aa = Replace(a0, ",", " ") 'Make all delimiters of same type, so removing commas, you may need to add more replace work here?
ab = Application.Trim(aa) 'Reduce all whitespace to single entries, i.e. " " rather than " "
ac = Split(ab, " ", -1) 'Now split by that single whitespace entry
Dim txt()
i2 = 0
lastIsNumeric = False
For i1 = 0 To UBound(ac) - 1 'Step through each entry in our "split" list
If IsNumeric(ac(i1)) = True And IsNumeric(ac(i1 + 1)) = True Then
'Two numbers back to back, join
ReDim Preserve txt(i2)
txt(i2) = ac(i1) + ac(i1 + 1)
i2 = i2 + 1
i1 = i1 + 1
Else
'Not two numbers back to back, don't join
ReDim Preserve txt(i2)
txt(i2) = ac(i1)
i2 = i2 + 1
End If
Next i1
If IsNumeric(ac(UBound(ac))) = False Then
'Need to add last entry to txt()
ReDim Preserve txt(UBound(txt) + 1)
txt(UBound(txt)) = ac(UBound(ac))
End If
End Sub
edit 2021-06-01:
The above will generate a list (txt) of all the entries within your address. You can then reassemble if you wish, or extract out the postcode only.
If you want it as a function, then it would be:
Public Function getPostcode(a0)
aa = Replace(a0, ",", " ")
ab = Application.Trim(aa)
ac = Split(ab, " ", -1)
Dim txt()
i2 = 0
lastIsNumeric = False
For i1 = 0 To UBound(ac) - 1
If IsNumeric(ac(i1)) = True And IsNumeric(ac(i1 + 1)) = True Then
'Two numbers back to back, join
ReDim Preserve txt(i2)
txt(i2) = ac(i1) + ac(i1 + 1)
i2 = i2 + 1
i1 = i1 + 1
Else
'Not two numbers back to back, don't join
ReDim Preserve txt(i2)
txt(i2) = ac(i1)
i2 = i2 + 1
End If
Next i1
If IsNumeric(ac(UBound(ac))) = False Then
'Need to add last entry to txt()
ReDim Preserve txt(UBound(txt) + 1)
txt(UBound(txt)) = ac(UBound(ac))
End If
'Re-assemble string for return
rtnTxt = ""
For i1 = 0 To UBound(txt)
rtnTxt = rtnTxt & " " & txt(i1)
Next i1
getPostcode = rtnTxt
End Function