正则表达式和检查出现

Question

I am actually dealing with regular expressions and i am still trying to understand how to approach properly this kind of problems.

So lets say i have this regular expression:

[A − Z]
∗01∗
[ˆ[A − Z]]{3}

On alphabet [AZ][0-9]

First question is: {3} means that there must be atleast 3 characters that belong to a "part" of the regular expression(lets say 3[A − Z]) or it is strictly refering to the last one ([ˆ[A − Z]])?

My second doubt is: if it is the last one, checking if there are atleast 3 occurrences might be easy(just 3 states that check if the char is a number, otherwise exit),right? Otherwise, if it might be any of the possible part of the regular expression, how do i check without a counter(eventually confirm if i shouldnt be using a counter) how many occurrences repeat in any possible state?

I am not really interested in a solution with code, i just want to fully understand the topic.

Answer 1

Regular expressions are a formal mathematical construction, but syntaxes for describing them may vary. In common syntaxes, {3} means the previous item is repeated three times. For example, [AB]{3} is the same as [AB][AB][AB] , so it will match AAA , AAB , ABA , ABA , BAA , BAB , BBA , or BBB . Or (AA|B){2} will match AAAA , AAB , BAA , or BB . It does not require there be two characters. It requires there be two matches of (AA|B) .

What the “previous item” is may depend on the particular syntax you are using. For example, in AA|B{2} , either | or {…} could be given a higher precedence, so it could be AA|(B{2}) or (AA|B){2} , depending on the rules in your syntax. However, in the specific example you asked about, the brackets clearly form a unit, so [ˆ[A − Z]]{3} requires three matches to [ˆ[A − Z]] . Again assuming a common syntax, [ˆ[A − Z]] means one character that does not match [AZ] , so a character that is not A through Z . Since your alphabet consists only of A through Z and 0 through 9 , [^[AZ]] matches 0 through 9 .

Thus [^[AZ]]{3} matches a three-digit numeral and nothing else.

Answer 2

First, there's a bunch of problems with your regex.

I believe your "smart" editor has mangled the regex. It's replaced ^ (U+0005E CIRCUMFLEX ACCENT) and - (U+0002D - HYPHEN-MINUS) with the fancy versions: ^ (U+002C6 - MODIFIER LETTER CIRCUMFLEX ACCENT) and (U+02212 - MINUS SIGN). They look the same, but they are different characters and have different meanings in a regex. To avoid this, be sure to use a good code editor such as Atom .

Spaces are also important. [A - Z] means something different than [AZ] . So are newlines, they are treated literally.

∗01∗ does not mean to match 01 surrounded by anything. Regexes don't work like file globs. While * does mean "zero or more" like a file glob, it is "zero or more of the immediately preceding thing". . matches (almost) anything. So it would be .*01.* .

[ˆ[A − Z]]{3} should be [^AZ]{3} . [^...] means to match what is not in the set. [^AZ]{3} means to match exactly 3 of anything which are not between A and Z. 123 or abc or !@# .

Putting it all together: [AZ].*01.*[^AZ]{3} says to match exactly one character in the set between A and Z, then match anything, then exactly 01 , then anything, then exactly 3 characters which are not in the set between A and Z. C01;;; and blah blah Z blah 01 blah blah abc both match.

Regex 101 is a valuable resource for understanding regexes. Regular-Expressions.info is a very good tutorial site.

First question is: {3} means that there must be atleast 3 characters that belong to a "part" of the regular expression(lets say 3[A − Z]) or it is strictly refering to the last one ([ˆ[A − Z]])?

{3} is a "quantifier" . So are + (one or more), * (zero or more), and ? (zero or one). All quantifiers match the thing immediately preceding it. A{3} means "AAA". [AZ]{3} means exactly three characters in the set of A through Z.

My second doubt is: if it is the last one, checking if there are atleast 3 occurrences might be easy(just 3 states that check if the char is a number, otherwise exit),right? Otherwise, if it might be any of the possible part of the regular expression, how do i check without a counter(eventually confirm if i shouldnt be using a counter) how many occurrences repeat in any possible state?

Regular expressions are insanely complicated. They are a language unto themselves. Unless this is for a class, use a regular expression library such as PCRE .