Bash - 13.1 - regex Expanded

Controlling Repeated Characters Through *, +, ?

These are used to express a variable number of times that preceding character in regex can appear in the string.

*: The preceding character can appear 0 or more times.
1*0* matches 11111000000, 111111110, 10, 0, or even an empty string.

+ : The preceding character can appear 1 or more times.
1+0+ matches 111110000, 1111110, 10
but not 0, 1111, or an empty string.

? : The preceding character can appear 0 or 1 time.
1?0? matches 1, 0, 10, or an empty string.

a*b*c* : Matches aaaaabbbc, abc, ab, bbbbb, or even empty string. (will match abacab).
a+b+c+ : Matches aaaaabbbc, abc, abbbbbc, but not an empty string or 1111.
a?b?c? : Matches abc, ab, or empty string.
None will match with abaaabcas the letters are not in correct sequence.

Literal Characters: A regex can contain literal characters without the metacharacters modifying them.
jpe?g : Matches jpeg or jpg (with e being optional and j,p, g a must).
hello!? : Matches hello or hello! (with ! being optional).

Using . (Dot)

. matches any single character except a newline.

b.t will match any string that has b followed by any one character, followed by t.

• It will match: bat, bbt, b0t, b#t, b t
• It will not match: bt (missing the middle character)
• batt (too many characters)

... (three dots) will match any three characters, including three blank spaces.

Combining with Other Metacharacters like *, +, and ? can modify . to allow for more flexible matches:

.*: Matches any sequence of characters (including zero characters). b.?t means b followed by zero or one character of any kind followed by a t.
b.*t will match: bat, bbt, b0t, b#t, b t, bt, batt, b t

.+: Matches any sequence of one or more characters.
b.+t will match: bat, bbbbbbt, b t, b?!!#&*t, bait, b00t
It will not match: bt (since there must be at least one character between b and t)

.?: Matches zero or one character of any kind.
b.?t will match: bt, bat, b?t, b#t, etc.

Controlling Where a Pattern Matches

A regular expression need not match the full string, it can match any substring within a string, which is a contiguous sequence of characters found within the larger string.

• a*b*c* will match abacab because it looks for zero or more occurrences of a, b, and c in any order.
• b.t will match batt as it matches b, any character (represented by .), and t. It will also match strings like bbbbbatttt because it matches the substring “bat”.
• b?.t and b?.t+ will also match substrings within bbbbbatttt.

regex patterns like a*, char*, or a*b* can be too broad, matching anything and everything, becoming useless.
To control where the regex matches, we can use anchors to enforce specific positions in the string.

• ^ : Asserts the beginning of a string.
• $ : Asserts the end of a string.
• ^ and $ : Asserts the start and end of the string, ensuring the entire string must match the pattern.

^b?.t:

• Will not match bbbbbbatttt because the pattern starts at the beginning of the string, where it expects an optional b (b?), followed by any character (.), and then t. However, the sequence does not match at the start.
• b? matches the first b, the . marks the second b but the t+ does not match as the next character is not t.

^b?.t+:

• This will not match bbbbbbatttt either because it starts with an optional b, followed by any character, and then t+ (one or more ts). The next character after the first b does not meet the pattern.
• Matches a string that starts with 0 or 1 b because b? insists that there should be no more than one b. The next should be any one character followed by t.

b?.t+$:

• This will match bbbbbatttt because it scans the string from left to right, allowing for zero or more bs, followed by one character, and ending with one or more ts at the end of the string.

^b?.t+$:

• This will match strings that start with zero or one b, followed by any character, and ending with one or more ts. batttt, bbttttt, attttt.

^0a*1+:

• This will match strings that start with a 0, followed by zero or more a’s, and at least one 1. 0a1, 0aaaa111, 101111, 01.
• There can be any character after 1. 0a1a 011100

^0a*1+$:

• This variation fixes the ending part to ensure it ends with one or more 1s, with no characters following the 1. It will match: 0a1, 0aaaa111, 10111.

^$:

• This matches the empty string. It asserts that there is nothing between the start and end of the string.

Matching From a List of Options

When we want to match a set of characters without caring about their specific order to indicate “any of these”, we can use character classes.
These are defined with square brackets ([]), which specify that the next character in the string must match any single character from the list of options inside the brackets.

Types of Lists: as given by the notation inside the bracket.

1. Enumerated List : [abcd] Matches any one of the characters a, b, c, or d, in any order. There should be no separators.
[abc][abc][abc] or [a-c][a-c][a-c]: This matches any three consecutive characters where each character is one of a, b, or c, in any combination.
abc, acb, bca, aaa, bbb, bcc, zyxaaa, @aaa@, aaaaa, 1cba2.
[abcxyz] [abcdABCD] [abc1234]

2. Range : [a-f] Matches any one of the characters from a to f, inclusive. This is equivalent to writing [a, b, c, d, e, f].
Range Combinations:

• [a-cx-z] : Matches any character from the sets a to c or x to z.
• [a-d1-4] : Matches any character from the sets a to d or 1 to 4.
• [a-dA-D] : Matches any character from the sets a to d or A to D.

3. Character Class : A POSIX character class is a predefined group that matches a set of characters based on a certain property.

• [:alpha:] matches any alphabetic character (equivalent to [a-zA-Z]).
• [:punct:] matches any punctuation character such as ., ,, !, etc. (there’s no direct range for all punctuation marks).