>>I am looking for a way to identify a 16 digit number preceeding by any non-digit character and ending with non-digit character or nothing. The pattern I defined seems to work. However, I need to replace only that 16 digit number with the 16X. I am not sure how to apply Replace command here.
>>
>>Thanks in advance.
>
> if your string is 'abc1234567890123456', ie end of line after the 16 digits, it won't match
> I modified the pattern so that the non digit after the 16 also matches the end of the string
>
>Also, note that I have inserted parentheses in the pattern, to group
>a) the non digit before the 16 digits
>b) the 16 digits
>c) the non digit after the 16 digits
>
>I will use those groups of parentheses below
>
>

>loRegExp = CREATEOBJECT("VBScript.Regexp")
> 
>loRegExp.PATTERN = "(\D)(\d{16})(\D|$)" && search for non digit following by 16 digits
> 
>loRegExp.IgnoreCase = .T.            && In this case, IgnoreCase = false is sufficient
> 
>loRegExp.GLOBAL = .T.
> 
>lcString = 'abc1234567890123456fd'
>lcString = 'abc1234567890123456'
>?m.loRegExp.Test(m.lcString)
>

>
>(1) you can use replace() - see below - see also page 19 first row http://www.atoutfox.org/modules/articles/pdf/0000000748.pdf
>
>

>&& With replace()
>
>&& if is optional
>if( m.loRegExp.Test(m.lcString) )
>	? 'old: ', m.lcString 
>	lcString = m.loRegExp.Replace(m.lcString, '$1$2X$3')
>	
>	? 'new: ', m.lcString 
>endif
>

>
>(2) Using Execute() is also possible and it gives you finer control
>
>Do not forget, however, to loop the collection backwards
>
>Using the SubMatches collection - see pages 8 and 9
>
>
>

>&& or with Execute()
>lcString = 'abc1234567890123456'
>
>matchesObj = m.loRegExp.Execute(m.lcString)
>
>? 'old: ', m.lcString 
>
>for i = m.matchesObj.Count - 1 to 0 step -1
>	match = m.matchesObj.item(m.i)
>	
>	
>	
>	?match.Submatches.Count
>	?match.Submatches[0]
>	?match.Submatches[1]
>	?match.Submatches[2]
>
>	s = m.match.Submatches[0] + m.match.Submatches[1] + 'X' + match.Submatches[2]
>	lcString = stuff(m.lcString, m.match.FirstIndex+1, m.match.Length, m.s)
>	
>	
>endfor
>? 'new: ', m.lcString 
>

Greg,

Thanks a lot. I've been reading yesterday your paper, this is how I figured the first part.