regex question - matching single whitepaces in range, but not multiple?

Caled

320gp      a copper-banded oak tun            Pestle & Mortar

This is just an example - but using multiple whitespaces as a boundary between capture groups is really really common and not just in muds if we're being honest. I feel that my way of doing it isn't 'right' even though it works. So I thought I'd ask if there is a better way.

Here are a couple of screenshots from regex101.com, the first illustrating the problem, and the second being how I get around it.

Sena

Like Rhivona said, you probably want to define a minimum or a maximum number of spaces, and a minimum and maximum number of other parts as well. But if you want to actually capture and use the item name in this particular example, without capturing a bunch of spaces after it, what you want is a non-greedy quantifier.  ^\d+gp\s+([a-zA-Z -]+)\s{2,}  will capture "a copper-banded oak tun          " as the item name for example, and you probably don't want all the extra whitespace at the end.

By default, each part of the pattern from left to right tries to match as much as it possibly can ("greedy"). If you add a ? after the quantifier (quantifiers are the +, *, {2,}, {1,5}, etc. that specify how many characters to match), it becomes non-greedy, matching as little as it possibly can. So if you do  ^\d+gp\s+([a-zA-Z -]+?)\s{2,}  then you'll correctly capture the item name as "a copper-banded oak tun".

There's a bigger problem with your method in this specific example though; there won't always be multiple spaces separating the item name from the shop name in directory, if the item name is long. For example:
1mc        a blown glass vial stamped with an The Shop of Wonders
In this case, your pattern won't work. There's still a way though, by specifying more minimum and maximum lengths for things. Here's a pattern I'd use to capture the price, item name, and shop name from directory:

^(\d+)(gp|cr|mc)\s+([a-z A-Z',()0-9&~"\.-]{34})\s+(.+)$

That {34} means to match exactly 34 characters, no more and no less. I also added other characters to the range ( [a-z A-Z',()0-9&~"\.-] ) that can appear in item names because I like to be as specific as possible, but .{34} will work just as well.
Unfortunately, you're back to capturing trailing spaces along with the item name using this pattern, because there's no clear marker that separates the item and shop name, but you can always remove trailing spaces in the code later. It is possible to change the pattern to avoid that, using lookarounds, but it's generally unnecessary. I can go into that later if you'd like.

Rhivona

something like this, maybe?

^\d+[g][p]\s{2,}[A-Za-z-\s]+\s{2,}.+$

\s{2,} makes it only note whitespaces in groups of two or more

Caled

Right. That's essentially what I was doing with \s\s though.

I was thinking there might be a way of changing this part:
[A-Za-z-\s]+

so that it will only ever match if the spaces inside are singular?

Sena

Caled said:

Right. That's essentially what I was doing with \s\s though.

I was thinking there might be a way of changing this part:
[A-Za-z-\s]+

so that it will only ever match if the spaces inside are singular?

Didn't see this while I was writing my post. It is possible to make that match only if it doesn't contain any double spaces, using a negative lookbehind for example:

[A-Za-z\s-]+(?<!\s{2})

The (?<!\s{2}) means to go back and check for \s{2}, and don't match if it's found. Keep in mind this will still match a single space at the end, so you'd get "an oaken vial " for example. And you still have to worry about the situation I mentioned where there aren't any double spaces.

Sena

Sorry for the triple post, but it's too late to edit. Another method of only matching single spaces in a phrase is to match any number of "word followed/preceded by a space". For example:

^\d+gp\s+((?: ?[a-zA-Z-]+)+)\s+

In your oak tun example, this will capture the item name with no extra spaces at the beginning or end. The first \s+ matches all the leading spaces, the (?: ?[a-zA-Z-]+)+ matches one or more instance of "zero or one spaces followed by a word", and the last \s+ will mark the end of the item name because it will match the first space that isn't followed by a word.

Caled

Thanks so much @Sena
The pattern in your first post solved my issue with directory matching. I hadn't seen the long descs when I posted, but I did soon after when I began looking for weapons. Using fixed lengths didn't occur to me. 

The other two solutions will come in handy for other things in the future (including outside mudding, at work.) So thanks for the really comprehensive answers!

Brenex

Great post! Always looking to practice and learn more about how to use regex effectively!