tcl：正则表达式匹配在字符串末尾

Question

I trying to match a substring occurs many times in string

str1 = st1.st2.{k}.st3.{k}.st4.{k}.
str2 = st1.st2.{k}.st3.{k}.st4.

I use regexp to match "{k}" at the end of str1:

regexp .*\.\{k\}\.$ $str1 

but I got 0 !! in fact I use regsub to test the regexp

regsub {.*\.\{k\}\.$} $str {}

result ==> empty

if the pattern is matched, the matched string will be removed !! what missing in regexp expression ?

Answer 1

In your code, the regexp is returning the value 1 only, not 0. When you want to match the last occurrence of .{k}. , you have to go ahead with sub-matches to get what you want.

set str1 st1.st2.{k}.st3.{k}.st4.{k}.
puts [regexp ".*(\.{k}\.)" $str1 whole last]
puts $last

Output :

1
.{k}.

The $ sign is not mandatory to specify the end of line as we simply want to match the last occurrence.

With the regsub , you should be using the back-reference to capture the 1st group, so that it can be replaced correctly.

puts [regsub "(.*)(\.{k}\.)" $str1 "\\1"] 

Output :

st1.st2.{k}.st3.{k}.st4

What is wrong with regsub {.*\\.\\{k\\}\\.$} $str {} ?

Well, the pattern .*\\.\\{k\\}\\.$ will match the whole string and you are replacing it with empty string, which is why you are getting the empty result.

Reference : Noncapturing Subpatterns