Regex question.

Posted by Blainer on Mon 18 May 2009 04:50 PM — 7 posts, 32,178 views.

Blainer Mon 18 May 2009 04:50 PM #0

When I run this code:

t = {}
s = "from=world, to=Lua"
for k, v in string.gmatch(s, "(%w+)=(%w+)") do
t[k] = v
end
for i,v in pairs(t) do Note(t[i]) end

I get output:
Lua
world

I am trying to learn regex so I can run this:

for v in io.lines([[speeds.txt]]) do
for k,w in string.gmatch(v,"pattern") do
-put each section into array
end
end

speeds.txt (I added the == trying to make some thing to match):

==1==Armory==Gueldar's Armory==2s3wn==
==2==Bakery==The Aylorian Eatery==2s3en==
==3==Bank==The Aylorian Bank of Ivar==11s==

I've been trying to regex for awhile now and noticed the output from the example above (taken from www) didn't seem to do what it was intended to do. So I figured I was missing something or using different syntax for patterns than MUSH.

David Haley USA Mon 18 May 2009 06:19 PM #1

What do you mean by not doing what it was supposed to do? Are you showing us your actual code? For instance you have your pattern as just "pattern" which won't do anything interesting.

Blainer Mon 18 May 2009 07:12 PM #2

The first bit of code doesn't seem to do what the programmer intended this made me think there was a difference with the way MUSH was using regex as compared with info on the web, it only list 2 of the 4 words. It was an example I was using to learn regex.

Example from web:
====
t = {}
s = "from=world, to=Lua"
for k, v in string.gmatch(s, "(%w+)=(%w+)") do
t[k] = v
end
for i,v in pairs(t) do Note(t[i]) end

I get output:
Lua
world
====

The second bit of code is my attempt to put a file into a table. I'm having troubles with the regex. I can't get it to match and noticed the code above did something that seemed wrong to me as I was trying to learn so I added it to post. "Pattern" would be a regex yes. I can't seem to get it to match any number of digits and any number of letters with spaces. I thought "=(%w*)=" would work but it doesn't.

This is the input file:
==1==Armory==Gueldar's Armory==2s3wn==
==2==Bakery==The Aylorian Eatery==2s3en==
==3==Bank==The Aylorian Bank of Ivar==11s==

I added the == to give my something to match.

This is what I have so far:
for v in io.lines([[speeds.txt]]) do
for k,w in string.gmatch(v,"pattern") do
-put each section into array
end
end

I am not even sure this is the best way to go about this but it seemed easier than string.find.

Nick Gammon Australia Forum Administrator Mon 18 May 2009 11:29 PM #3

Yes, but the word "pattern" won't match, as the word "pattern" doesn't appear in the string. You need to put in something that is likely to work, or we can't help you. For example:


for v in io.lines("speeds.txt") do
  for k, w in string.gmatch(v,"(%w+)=(%w+)") do
  --put each section into array
  end
end

Even then this won't work on your test data, as you have == between each word, and the regexp has a single = sign. I am not sure you are understanding what a pattern is. You don't use the word "pattern", you have to make one that is relevant to what you are trying to match.

This page shows what you can use in a pattern:

http://www.gammon.com.au/scripts/doc.php?lua=string.find

The regexp patterns are the same for string.find as string.gmatch, they just work a bit differently in how they return the matching data.

Blainer Tue 19 May 2009 03:03 AM #4

Sorry guys I've made these posts as clear as mud. I am asking two questions.

Question 1.
The code I am trying to make work is:

for v in io.lines([[c:\speeds.txt]]) do
for k,w in string.gmatch(v,"=(%w*)=") do
Note(w)
end
end

speeds.txt looks like this:
==1==Armory==Gueldar's Armory==2s3wn==
==2==Bakery==The Aylorian Eatery==2s3en==
==3==Bank==The Aylorian Bank of Ivar==11s==

The regex above matches the whole line and I can't figure out how to get it to match one part of a line at a time. I need to break speeds.txt into four parts per line, Number, Name, Disc, Speedwalk. I can't get my head around this. "=" is an alphanumeric so it matches the whole line, is there a better character I should use to mark the sections of speeds.txt. Or can a regex match each part of the line with no character marking the beginning and end of each section?

I have tried [] ranges to avoid matching the whole line and even tried using the exact words "==(1)==(Armory)==(Gueldar's Armory)==(2s3wn)==" to test and still can't match each section. As I understand gmatch should should match 4 times on the first line using "==(1)==(Armory)==(Gueldar's Armory)==(2s3wn)==" and iterate 4 times returning what is between the brackets. How ever when returns "Armory" twice when I try this.

Question 2.
If I run:

t = {}
s = "from=world, to=Lua"
for k, v in string.gmatch(s, "(%w+)=(%w+)") do
t[k] = v
end
for i,v in pairs(t) do Note(t[i]) end

And I get output:
Lua
world

Is this code working correctly. As I understand gmatch with regex it should give the out put:
from
world
to
Lua

Sorry for all the confusion and thanks for the help.

Nick Gammon Australia Forum Administrator Tue 19 May 2009 04:06 AM #5

Question 1.

The string.gmatch iterator returns the next match from the line. This is a single value (as you had a single capture), so for it to work you need something like this:


for v in io.lines("speeds.txt") do
  for w in string.gmatch(v,"=(%w+)=") do
    Note(w)
  end
end


Output

1
Armory
2s3wn
2
Bakery
2s3en
3
Bank
11s

Question 2.

You are capturing the information OK, but your debugging display is wrong - it only shows the value, not the key. You could fix that up, but I used tprint to show what was happening:


t = {}
s = "from=world, to=Lua"
for k, v in string.gmatch(s, "(%w+)=(%w+)") do
t[k] = v
end

require "tprint"

tprint (t)

Output

"to"="Lua"
"from"="world"

Blainer Tue 19 May 2009 05:16 AM #6

Ah great thanks Nick. I suspected my regex was correct.