PCRE Regex help

Posted by Madcowchow on Mon 06 Nov 2017 02:00 PM — 5 posts, 24,615 views.

Madcowchow Kenya Mon 06 Nov 2017 02:00 PM #0

Hello Friends

Please point me to the right direction how I can extract the table with a regex as below.


local mystring = "You are madcowchow (dark-elf, archer) aged 195 years."
local re = rex.new("^You are (?<name>\w+) \((?<tribe>\S+), (?<job>\w+)\) aged (?<age>\d+) years\.$")
local s, e, t = re:match(mystring)
print (type(t))

edit: i am expecting output to return 'table', but get 'nil'.
the regex extracts the fields when i test it at https://regex101.com/

Thank you

Nick Gammon Australia Forum Administrator Mon 06 Nov 2017 08:51 PM #1

Your literal strings have backslashes in them, which Lua processes before it even gets to the regexp. Thus you have to double them, like below:


local mystring = "You are madcowchow (dark-elf, archer) aged 195 years."
local re = rex.new("^You are (?<name>\\w+) \\((?<tribe>\\S+), (?<job>\\w+)\\) aged (?<age>\\d+) years\\.$")
local s, e, t = re:match(mystring)
print (type(t))

require "tprint"
tprint (t)

Result:


table
1="madcowchow"
2="dark-elf"
3="archer"
4="195"
"tribe"="dark-elf"
"name"="madcowchow"
"age"="195"
"job"="archer"

This is probably why Lua uses "%" in its pattern matching rather than "\" to save the trouble of doing that.

Madcowchow Kenya Mon 06 Nov 2017 09:34 PM #2

Thank you Nick, much appreciated. Spent most of my morning trying to see where I failed. I guess you tend to forget stuff as u grow older :)

Marco Tue 07 Nov 2017 11:49 AM #3

Sorry if i ask, but i'm very interested in this topic too.

Looking at the code and understanding what you've said (Nick) :


local mystring = "You are madcowchow (dark-elf, archer) aged 195 years."
local re = rex.new("^You are (?<name>\\w+) \\((?<tribe>\\S+), (?<job>\\w+)\\) aged (?<age>\\d+) years\\.$")
local s, e, t = re:match(mystring)
print (type(t))

require "tprint"
tprint (t)

Result:


table
1="madcowchow"
2="dark-elf"
3="archer"
4="195"
"tribe"="dark-elf"
"name"="madcowchow"
"age"="195"
"job"="archer"

Is it normal that if i put something similar in a Trigger (or even an alias) and in the SendTo Script it doesn't work?

It is supposed to work correctly only in a Script file?

Nick Gammon Australia Forum Administrator Tue 07 Nov 2017 08:06 PM #4

The problem with triggers and aliases (using the Send box) is that the client expands out backslashes first, then sends it to Lua which expands backslashes again. Thus in Send To Script you have to double the backslashes again (that is, four backslashes for every one you really want).

This is obviously fairly tedious. I suggest using a script file, or using a Lua pattern (eg. string.match) which avoids the problem because it uses "%" to escape special fields (like alphabetic sequences) instead of "\".