CODEWITHSUNDEEP
ruby-on-railsrubyregex
Sampat Badhe
Sampat Badhe

Reputation: 9095

regex to remove characters from the string

I have street address and I want to replace the street direction with blank string.

this are the streets E, W, S, N, NE, NW, SE, SW which I want to replace with blank string

I have tried this regex /((e|w)|(n|s)(e|w)?)/i to replace the direction, but this replacing all the instance characters present in regex.

I would like convert street address like following - 

"123 Main Street S" => "123 Main Street"
"E 123 Main Street" => "123 Main Street"
"SE 123 Main Street" => "123 Main Street"
"123 Main Street, SW, US" => "123 Main Street, US"
"123 Main Street, s, USA" => "123 Main Street, USA"

Upvotes: 1

Views: 313

Answers (3)

Wiktor Stribiżew
Wiktor Stribiżew

Reputation: 627488

Here is the best I could come up with:

/[\s,]*\b(?:[EW]|[NS][EW]?)\b/i

The regex matches:

  • [\s,]* - zero or more whitespaces and commas
  • \b - word boundary (leading)
  • (?:[EW]|[NS][EW]?) - a non-capturing grouping construct that matches:
    • [EW] - either E or W
    • [NS][EW]? - either N or S optionally (thus these are missing from the first alternative) followed by either E or W.
  • \b - trailing word boundary

NOTE that using character classes (those [...] thingies) to match single characters is the best practice as they do not cause backtracking (so it is preferred to groups with alternatives that are usually used with sequences of characters).

See IDEONE demo

rx = /[\s,]*\b(?:[EW]|[NS][EW]?)\b/i
puts "123 Main Street S".sub(rx, '').strip  # => "123 Main Street"
puts "E 123 Main Street".sub(rx, '').strip  # => "123 Main Street"
puts "SE 123 Main Street".sub(rx, '').strip # => "123 Main Street"
puts "123 Main Street, SW, US".sub(rx, '').strip # => "123 Main Street, US"

Upvotes: 3

zsltg
zsltg

Reputation: 773

/,?\s?\b((e|w)|(n|s)(e|w)?)\b\s?/i

This would also remove the blank spaces around the directions and preceding comma.

Example: https://regex101.com/r/iW8dZ7/2

Upvotes: 1

vks
vks

Reputation: 67988

\b((e|w)|(n|s)(e|w)?)\b

Should do it for you.\b is word boundary.See demo.

https://regex101.com/r/iJ7bT6/8

Upvotes: 3

Related Questions