Regsubbing simple matches

Question

I'm looking for a regsub example that does the following:

123tcl456TCL789 => 123!tcl!456!TCL!789
This is an Tcl example => This is an !Tcl! example

Yes, I could use string first to find a position and mash things but I saw in past a regsub command that does what I want but I can't recall. What would be the regsub command that allows that? I would guess regsub -all -nocase is a start. I am bad at regsub and regexps. I wonder if there is a site or tool/script that we can supply a string, the final result and then we get the regsub form.

Answer 1

You're looking at the right tool, but there are various options, depending on exactly what the conditions are when faced with other text. Here's one that wraps each occurrence of "Tcl" (any capitalisation) with exclamation marks:

set inputString "123tcl456TCL789"
set replaced [regsub -all -nocase {tcl} $inputString {!&!}]
puts $replaced

That's using a very simple regular expression with the -nocase option, and the replacement means "put ! on either side of the substring matched".

Another (more generally applicable... perhaps) might be to put ! after any letter or number sequence that is followed by a number or letter.

set replaced [regsub -all {[A-Za-z]+(?=[0-9])|[0-9]+(?=[A-Za-z])} $inputString {&!}]

Note that doing things correctly typically requires understanding the real input data fairly well. For example, whether the numbers include floating point numbers in scientific notation, or whether the substrings to delimit are of fixed length.