Error when substituting a backslash in a string

Ruby
1

Hi,

Why can’t I substitute a single backslash in a string?

a = “\6”
p a # => “\6”
puts a # => \6
p a.sub(’’, ‘’)

Expected: “\6”
Got:
j:4: unterminated string meets end of file
j:4: parse error, unexpected $, expecting ‘)’

If I change the replacement to ‘c’, I get a new error:

j:4: parse error, unexpected tIDENTIFIER, expecting ‘)’
puts a.sub(’’, ‘c’)
^
j:4: unterminated string meets end of file

Both substitution work fine when the pattern is a regular character.
The substitution also works if num(’’) is a multiple of 2, example:
a = ‘\6’
p a # => “\6”
puts a # => \6
p a.sub(’\’, ‘’)

Expected: “6”
Got: “6”

Thanks for your time.

2

On 6/6/07, Federico Zagarzazú [email protected] wrote:

Hi,

Why can’t I substitute a single backslash in a string?

a = “\6”
p a # => “\6”
puts a # => \6
p a.sub('', ‘’)

On that last line, your backslash is escaping the single quote and
causing parsing lossage.

How’s about this:
p a.sub(“\”, ‘’)

Hope that helps,
-Harold

3

On Jun 6, 2:59 pm, “Federico Zagarzazú” [email protected] wrote:

Why can’t I substitute a single backslash in a string?
a = “\6”
…
p a.sub('', ‘’)

Because you’re essentially escaping a single quote in your String#sub
method call. Double escape that too:

irb(main):001:0> a = “\6”
=> “\6”
irb(main):002:0> a.sub(‘\’, ‘’)
=> “6”

Why is this? Suppose I wanted to make a string with single quotes in
it:

irb(main):003:0> a = ‘'hello'’
=> “‘hello’”

4

Thanks for your answers, silly me, got confused :),

But, how does it work?, suppouse you want to convert “fede” to “f\4de”

  1. puts ‘fede’.sub(‘e’, ‘\4’) # => fde
  2. puts ‘fede’.sub(‘e’, ‘\4’) # => fde
  3. puts ‘fede’.sub(‘e’, ‘\\4’) # => f\4de
  4. puts ‘fede’.sub(‘e’, ‘\\4’) # => f\4de
  5. puts ‘fede’.sub(‘e’, ‘\\\4’) # => f\de
  6. puts ‘fede’.sub(‘e’, ‘\\\4’) # => f\de
  7. puts ‘fede’.sub(‘e’, ‘\\\\4’) # => f\4de
    …
    I don’t get it…

I was thinking that it might work by making two passes looking for and
converting backslashes, like this:
(7):
first pass : (\)(\)(\)(\4)
\ \ \ \4
second pass: (\)(\)4
result : \4

But I’m not 100% sure, do you know how it works?

Thanks again.

5

On Jun 6, 4:27 pm, “Federico Zagarzazú” [email protected] wrote:

  1. puts ‘fede’.sub(‘e’, ‘\4’) # => fde
  2. puts ‘fede’.sub(‘e’, ‘\4’) # => fde
  3. puts ‘fede’.sub(‘e’, ‘\\4’) # => f\4de
  4. puts ‘fede’.sub(‘e’, ‘\\4’) # => f\4de
  5. puts ‘fede’.sub(‘e’, ‘\\\4’) # => f\de
  6. puts ‘fede’.sub(‘e’, ‘\\\4’) # => f\de
  7. puts ‘fede’.sub(‘e’, ‘\\\\4’) # => f\4de

You have to be careful with using backslashes with a number for a
replacement string. In a replacement string, # (where # is a number)
indicates that you want to insert something that was matched in the
original string - the matched things are in parentheses and occur in
regular expressions. For example:

irb(main):001:0> a = “Page 3 of 86”
=> “Page 3 of 86”
irb(main):002:0> a.sub(/^Page (\d+) of (\d+)$/, ‘You are on page \1
out of a total of \2 pages’)
=> “You are on page 3 out of a total of 86 pages”

The PickAxe has an excellent subject on this and goes over your
question with all sorts of backslash permutations. I guarantee that
it will clarify things further for you:

http://www.rubycentral.com/book/tut_stdtypes.html
(see “Backslash Sequences in the Substitution” section)

6

Thanks for the link, it cleared my doubts.

7

p a.sub(’’, ‘’)

On that last line, your backslash is escaping the single quote and
causing parsing lossage.

so… sometimes escaping happens within single quotes, and sometimes it
doesn’t?

a = ‘yo\n’ # doesn’t escape

feels like this should escape:
p a.sub("\", ‘’)

and this shouldn’t:
p a.sub(’\’, ‘’)

one more thing to remember I guess… :slight_smile:

8

From: http://www.rubycentral.com/book/tut_stdtypes.html

“Within single-quoted strings, two consecutive backslashes are replaced
by a single backslash, and a backslash followed by a single quote
becomes a single quote. Double-quoted strings support a boatload more
escape sequences.”

easy enough :slight_smile: