[Issue 5502] New: More handy ways to create associative arrays

[d-bugmail at puremagic.com]

http://d.puremagic.com/issues/show_bug.cgi?id=5502

           Summary: More handy ways to create associative arrays
           Product: D
           Version: D2
          Platform: All
        OS/Version: Windows
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: Phobos
        AssignedTo: nobody at puremagic.com
        ReportedBy: bearophile_hugs at eml.cc


--- Comment #0 from bearophile_hugs at eml.cc 2011-01-29 10:11:56 PST ---
I think some helpers to create built-in associative arrays will be useful. Such
helpers may be built-in methods of the AAs, or free functions in std.algorithm,
or a mix of the two.

Three of the most common ways to build an associative array (examples in Python
2.6.6):

1) One of the most useful ways to create an AA is from a sequence of pairs. The
pairs may come from zipping of two iterables:

>>> a
[1, 2, 3, 4, 5, 6, 7, 8, 9]
>>> b = range(10, 100, 10)
>>> b
[10, 20, 30, 40, 50, 60, 70, 80, 90]
>>> zip(a, b)
[(1, 10), (2, 20), (3, 30), (4, 40), (5, 50), (6, 60), (7, 70), (8, 80), (9,
90)]
>>> dict(zip(a, b))
{1: 10, 2: 20, 3: 30, 4: 40, 5: 50, 6: 60, 7: 70, 8: 80, 9: 90}


Or from sorting and then grouping:

>>> from itertools import groupby
>>> s = "abaacaabbabaa"
>>> t = "".join(sorted(s))
>>> t
'aaaaaaaabbbbc'
>>> [(h, len(list(g))) for h,g in groupby(t)]
[('a', 8), ('b', 4), ('c', 1)]
>>> dict((h, len(list(g))) for h,g in groupby(t))
{'a': 8, 'c': 1, 'b': 4}


In Phobos of DMD 2.051 there are a zip() and group() that return such iterables
of pairs. So the only needed thing is a free function to build an AA from such
sequences, or better a built-in AA method like AA.fromPairs().

See also the Scala xs.toMap method:
http://www.scala-lang.org/api/current/scala/collection/immutable/List.html


See also AA.byPair() from bug 5466. AA.fromPairs(AA.byPair()) is an identity.

----------------------

2) Often you need to create an associative array that represents key-frequency
pair, but you want to avoid to waste time sorting the items and then grouping
them:


>>> from random import randint
>>> from collections import defaultdict
>>> d = defaultdict(int)
>>> s = "abaacaabbabaa"
>>> for c in s: d[c] += 1
...
>>> d
defaultdict(<type 'int'>, {'a': 8, 'c': 1, 'b': 4})


This is equivalent to a way to create a bag data structure.
For this purpose I don't suggest to add a built-in method to AAs. A free
frequency() function for std.algorithm will be enough.

----------------------

3) A simple and common way to create an associative array is from an iterable
of keys, with constant values:


>>> a = [1, 2, 3, 4, 5, 6, 7, 8, 9]
>>> dict.fromkeys(a)
{1: None, 2: None, 3: None, 4: None, 5: None, 6: None, 7: None, 8: None, 9:
None}
>>> dict.fromkeys(a, 1)
{1: 1, 2: 1, 3: 1, 4: 1, 5: 1, 6: 1, 7: 1, 8: 1, 9: 1}


This may become a built-in AA method like AA.fromKeys(), or a similar free
function in std.algorithm (in a statically typed language as D the type of the
values must be somehow given or be known, even when fromKeys() is allowed to
use a default value).

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------