Learn Regex

The original site: regexlearn

Test

If you can answer these, you don’t need to go through this course:

1. What is this ^\w+\.pdf$?
2. How to select the year in Release 10/9/2021?

Learn

Character Sets (All, numbers, alphabets, negations)

• we can directly search for the string of characters verbatim, we are looking for
• the period (dot) . is for selecting any kind of character, including special chars and spaces (full stop for any character)
  • 
• we can use character sets [abc]. This basically means c[au]t will select cat and cut. The alternative characters, a or u will be checked with each possibility.
  • b[aeiou]r selects all of bar, ber…, bur etc
• we can use negated character sets [^abc] to ensure that our selections will not have a and b and c in that particular spot
  • b[^eo]r selects bar, bir and bur
• we can use alphanumeric range (based on ASCII), [a-z] or [0-9]
  • [e-o] selects all alphabets efghijklmno
  • [3-6] selects 3456

Repetition

• these are placed after the characters

• +: 0 or 1(+) times

  • be*r → br, ber, beer etc

• *: 1(+) times

  • be+r → ber, beer (not br, doesn’t get selected)

• ?: optional (single) character

  • colou?r → color and colour (u is optional)

• {n} to show how many times, the character it entails should repeat

  • be{2}r → beer

• {n,} at least $n$ times.

  • be{2,}r → beer, beeer, beeer

• {n,m} at least $n$ times and at most $m$ times.

  • be{1,3}r → ber, beer, beeer

• [0-9]{4} selects the year in 10/9/2004

Grouping and Piping

This is where things get fun (and complex!)

• Parenthesis: (abc) to group abc
• We can reference an already mentioned group by writing \n where $n$ can be any integer, e.g. \1, \2 etc
  • (ha)-\1,(haa)-\2: is equivalent to writing (ha)-(ha),(haa)-(haa)
  • so, \1 references (ha) and \2 references (haa)
• Non-Capturing Grouping: (?:) to ensure that it is not captured by references.
  • In the previous example, (?:ha)-(haa)-\1-\1 gives, (ha)-(haa)-(haa)-(haa)
  • notice how (ha) is avoided from being referenced due to that (?:) syntax.
• Pipe Character |
  • (c|r)at|dog selects cat, rat and dot.
  • Compare it with [abc] that operates at a character level, the | character operates at an expression level

Escape and special sequences

• Escape Character \
• Caret Sign: ^ Selecting the line start
  • Dollar Sign: $ Selecting by end of the line
• Word Character \w: Any letter, number and underscore
  • Except Word Character \W: Except letters, numbers and underscore
• Digits: \d (only numbers)
  • Except Digits: \D (except numbers)
• Space: \s only spaces
  • Except Space: \S only spaces

Lookarounds

• Positive Lookahead (?=)
  • \d+(?=PM) looks for the number 3 in the expression Date: 4 Aug 3PM
  • the digit follows PM immediately
• Negative Lookahead (?!)¹
  • \d+(?!PM) looks for the number 4 in the expression Date: 4 Aug 3PM
• Positive Lookbehind (?<=)
  • (?<=\$)\d+ looks for the number 5 in the expression `Product Code: 1064 Price: $5
• Negative Lookbehind (?<!)
  • (?<!\$)\d+ looks for the number 1064 in the expression `Product Code: 1064 Price: $5

Flags

• a.k.a. modifiers determine whether typed expressions
  • treat text as separate lines
  • are case sensitive
  • find all matches
• global flag: /<your regex expression>/g
  • select all matches, (if not, then only the first one)
• multiline flag: /<rgx>/m
  • regex sees all text as one line
  • enabling this flag, makes regex respect the newlines
  • makes the $ point to the last char of the first line
  • /<rgx>/gm will select all in each line
• case insensitive flag: /<rgx>/i
  • don’t worry about case bro

Concept

• greedy matching: matchmaking will be as long as possible
  • Doesn’t stop at ber… goes till beeeer
• lazy matching: stops at the first matching
  • Stops at ber. lazy chico

Done!

Footnotes

1. instead of = in positive lookahead, we put the negation symbol ! ↩