sorteddict PEP proposal [started off as orderedict]

[Mark Summerfield]

On 2007-09-25, Andrew Durdin wrote:
> On 9/25/07, Mark Summerfield <m.n.summerfield at googlemail.com> wrote:
> >     Since the sorteddict's data is always kept in key order, indexes
> >     (integer offsets) into the sorteddict make sense.  Five additional
> >     methods are proposed to take advantage of this:
> >
> >     key(index : int) -> value
> >
> >     item(index : int) -> (key, value)
> >
> >     value(index : int) -> key
> >
> >     set_value(index : int, value)
> >
> >     delete(index : int)
>
> But what about using non-sequential integer keys (something I do quite
> often)?
>
> e.g. sorteddict({1:'a', 3:'b': 5:'c', 99:'d'})[3]  should return 'b', not
> 'd'.
>
> Andrew

Hmmm, managed to confuse myself with 'b' and 'd'!

d = sorteddict({1:"one", 3:"three", 5:"five", 99:"ninetynine"})

d.items()

[(1, 'one'), (3, 'three'), (5, 'five'), (99, 'ninetynine')]

d[3], d.value(3)

('three', 'ninetynine')

So using [] returns the value for the given key and using value()
returns the value for the given index position.

-- 
Mark Summerfield, Qtrac Ltd., www.qtrac.eu