Let "".split() split according to _any_ whitespace

[kseistrup]

Python's "".split() method has a useful feature: When invoked without any arguments, the string is split according to any whitespace, and empty strings are discarded from the result:

>>> s = "X Y\tZ"
>>> s.split(" ")
['X', 'Y\tZ']
>>> s.split("\t")
['X Y', 'Z']
>>> s.split()
['X', 'Y', 'Z']

Having a similar logic in ABS would make it much easier to split strings without having to resort to calling .trim() on the resulting elements.

[odino]

On the fence on this. JavaScript uses "" as default when no argument is passed, so I was thinking of doing the same, but I'm open to suggestion -- especially considering I'm more inclined to lean towards what Python does in general over other languages :)

What are the characters that are considered "empty strings"? I see \t in the example, had a look at the doc but it's not clear to me. EDIT: found it.

[kseistrup]

Caveat: I'm not a big fan of JavaScript, so I'm biased on this issue.

From a practical standpoint, if I wanted to do something with every character in a string in Python, I would use

for c in "xyz":
    do_something(c)

or

chars = list("xyz")

Also from a practical standpoint: If ABS were to use JavaScript's understanding of the default, in order to split a simple string like "X Y\tZ" I would have to loop over the string and its substrings with multiple calls to .split() and .trim(), which seems a bit over the top.

Also, if we look how Bash does it, it would split an unquoted environment variable at the characters in $IFS, that defaults to " \t\n":

s=$(printf 'X Y\tZ')
for c in $s; do echo $c; done
X
Y
Z

---

PS: If ABS decides to go with Python or Bash's strategy, ABS should possibly consider letting .trim() without any arguments follow the same logic. E.g., in Python:

>>> '\t \nX\f\v'.strip()
'X'

[kseistrup]

PPS: I just tried JavaScript:

$ js52
js> "XYZ".split()
["XYZ"]
js> "XYZ".split("")
["X", "Y", "Z"]
js> ^D
$ node
> "XYZ".split()
[ 'XYZ' ]
> "XYZ".split("")
[ 'X', 'Y', 'Z' ]
> 

[odino]

You won me with the for in example :) I think it makes sense to do for char in str.split()

…

On Thu, Dec 27, 2018 at 3:18 PM Klaus Alexander Seistrup < ***@***.***> wrote: Caveat: I'm not a big fan of JavaScript, so I'm biased on this issue. From a practical standpoint, if I wanted to do something with every character in a string in Python, I would use for c in "xyz": do_something(c) or chars = list("xyz") Also from a practical standpoint: If ABS were to use JavaScript's understanding of the default, in order to split a simple string like "X Y\tZ" I would have to loop over the string and its substrings with multiple calls to .split() and .trim(), which seems a bit over the top. Also, if we look how Bash does it, it would split an unquoted environment variable at the characters in $IFS, that defaults to " \t\n": s=$(printf 'X Y\tZ')for c in $s; do echo $c; done X Y Z ------------------------------ PS: If ABS decides to go with Python or Bash's strategy, ABS should possibly consider letting .trim() without any arguments follow the same logic. E.g., in Python: >>> '\t \nX\f\v'.strip()'X' — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#86 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAUC5Cx0h_XXhRCJeLs5WRut6YPMT_35ks5u9Kx8gaJpZM4ZijnX> .

-- Alex Nadalin http://www.linkedin.com/in/alessandronadalin www.odino.org www.twitter.com/_odino_