Project

General

Profile

Actions

Feature #9116

open

String#rsplit missing

Added by artagnon (Ramkumar Ramachandra) over 10 years ago. Updated almost 7 years ago.

Status:
Open
Assignee:
-
Target version:
-
[ruby-core:58375]

Description

There's nothing corresponding to Python's rsplit(). A quick glance at rb_str_split_m() tells me that it should be pretty trivial to implement. Is there any specific reason it hasn't already been done?

Updated by phluid61 (Matthew Kerwin) over 10 years ago

On Nov 16, 2013 6:35 PM, "artagnon (Ramkumar Ramachandra)" <
> wrote:

There's nothing corresponding to Python's rsplit(). A quick glance at
rb_str_split_m() tells me that it should be pretty trivial to implement. Is
there any specific reason it hasn't already been done?

What is rsplit? How does it differ from split (with or without a second
paramater)? What is the use-case/demand for it?

Sent from my phone, so excuse the typos.

Updated by shevegen (Robert A. Heiler) over 10 years ago

I am still not sure how it differs from #split().

Updated by shevegen (Robert A. Heiler) over 10 years ago

Oh, now I see:

"Except for splitting from the right, rsplit() behaves like split() which is described in detail below."

So it basically splits from the right, not left, unlike #split() which splits from the left, I think.

I suppose in Ruby you can do .reverse.split, but perhaps rsplit may be more convenient. (I myself don't think I have had a need to use something similar to rsplit so far).

Updated by phluid61 (Matthew Kerwin) over 10 years ago

I, too, looked up and read the documentation, a couple of times.

I understand that the difference only applies when a limit parameter is given, and so examples of the new API would be:

'a.b.c'.rsplit('.')    #=> ["a", "b", "c"], same as #split
'a.b.c'.rpslit('.', 2) #=> ["a.b", "c"]

I would want to clarify some of the other edge cases (from String#split) before continuing:

  • If pattern is a String, then its contents are used as the delimiter when splitting str. If pattern is a single space, str is split on whitespace, with leading whitespace and runs of contiguous whitespace characters ignored.

Would this have some right-handed equivalent in #rsplit? E.g. "...with trailing whitespace and runs..."? Or would it remain the same as #split? Or some third option?

E.g.:

' x  y '.rsplit(' ')   #=> ["x", "y"], same as split?
' x  y '.split(' ',-1) #=> ["", "x", "y"] or ["x", "y", ""] or ..?
  • If the limit parameter is omitted, trailing null fields are suppressed. If limit is a positive number, at most that number of fields will be returned (if limit is 1, the entire string is returned as the only entry in an array). If negative, there is no limit to the number of fields returned, and trailing null fields are not suppressed.

Similarly, would this become: "...leading null fields..." in both instances?

E.g.:

'..x..'.rsplit('.')    #=> ["x", "", ""] or ["", "", "x"] or ..?
'..x..'.rsplit('.',-1) #=> ["", "", "x", "", ""], same as #split?

Note that this would be another difference from #split, which doesn't depend on the limit parameter.

Seems like a lot of work. What is the demand for this feature?

Updated by alexeymuranov (Alexey Muranov) over 10 years ago

phluid61 (Matthew Kerwin) wrote:

I understand that the difference only applies when a limit parameter is given

It is not only when limit parameter is given:

"aaa".split("aa")  # => ["", "a"]
"aaa".rsplit("aa") # => ["a", ""]

Maybe with a regex there can be a more meaningful example.

Updated by alexeymuranov (Alexey Muranov) over 10 years ago

phluid61 (Matthew Kerwin) wrote:

Would this have some right-handed equivalent in #rsplit? E.g. "...with trailing whitespace and runs..."? Or would it remain the same as #split? Or some third option?

IMO, if it is introduced, i would say it must be completely symmetric with split.

Updated by phluid61 (Matthew Kerwin) over 10 years ago

alexeymuranov (Alexey Muranov) wrote:

It is not only when limit parameter is given:

"aaa".split("aa")  # => ["", "a"]
"aaa".rsplit("aa") # => ["a", ""]

Ah, I see. Thank you.

Maybe with a regex there can be a more meaningful example.

I'm interested to see how it would be achieved with regex, from an implementation point of view.

Updated by stomar (Marcus Stollsteimer) about 7 years ago

I'd like to revive the discussion about String#rsplit.

Here one use case I stumbled upon recently: splitting the digest off the end of a cookie (taken from Rack::Session::Cookie, see https://github.com/rack/rack/blob/master/lib/rack/session/cookie.rb#L139 ):

session_data = "session--data--digest"

digest, session_data = session_data.reverse.split("--", 2)
digest.reverse!        if digest
session_data.reverse!  if session_data

session_data  # => "session--data"
digest        # => "digest"

Note that each substring needs to be reversed back (for higher limits this would probably be done using #map), which seems inefficient and unhandy.

With rsplit this would become:

session_data = "session--data--digest"

session_data, digest = session_data.rsplit("--", 2)

Updated by shyouhei (Shyouhei Urabe) about 7 years ago

We looked at this issue in today's developer meeting.

Someone there pointed out that in general, rsplit(..., 2) could be avoided by using regexp, because you can match $ (for instance, you can match /(.+)--([^-]+)$/ for the session cookie). So the cookie case is a bit short for us to add new method.

It might make sense when the limit is bigger than 2. Is there such example?

Updated by naruse (Yui NARUSE) almost 7 years ago

I find a use case.

test_priv=# SELECT * FROM pg_database;
       datname        | datdba | encoding | datcollate  |  datctype   | datistemplate | datallowconn | datconnlimit | datlastsysoid | datfrozenxid | datminmxid | dattablespace |                 datacl
----------------------+--------+----------+-------------+-------------+---------------+--------------+--------------+---------------+--------------+------------+---------------+-----------------------------------------
 test_prev            |  16384 |        6 | en_US.UTF-8 | en_US.UTF-8 | f             | t            |           -1 |         12668 |         1917 |          1 |          1663 | {=Tc/pg,pg=CTc/pg,"\"foo=bar\"=CTc/pg"}

On PostgreSQL, I can retrieve databases information by SQL, and its ACL data is "datacl" column as key-value.
Its key is rolename (user/group name) and the value is the privilege, with separated "=".
The rolename may include "=" as above.

Actions

Also available in: Atom PDF

Like0
Like0Like0Like0Like0Like0Like0Like0Like0Like0Like0Like0