Any one up for an algorithm challenge?
I coded my solution to it yesterday - and it's basic hack work. I
am sure there is a more elegant way - and probably several :-)
THE SET UP
We have a set of data, represented by a series of objects in a
container block:
data: reduce [
make object! [id: "dd" items: [1 2 3 4 99] ]
make object! [id: "aa" items: [1 5 89 13] ]
make object! [id: "xx" items: [] ]
make object! [id: "yy" items: [6 5 4 3 2] ]
]
What we want is a subset range of these.... a sort of analog of
skip copy/part ....
But that only works if the data is in one block - ours is spread
across multiple blocks
EXAMPLES
Assume we (you!) have written a function call get-subset that
takes two args:
1. the data block
2. the range block: a block of two numbers being
the start and end values of the items we want to extract.
These are some sample expected results:
get-subset data [1 6] ;; ie items 1 through 6
[
make object! [id: "dd" items: [1 2 3 4 99] ]
make object! [id: "aa" items: [1] ]
]
get-subset data [5 11] ;; items 5 through 11
[
make object! [id: "dd" items: [99] ]
make object! [id: "aa" items: [1 5 89 13] ]
make object! [id: "yy" items: [6 5] ]
]
get-subset data [7 7] ;; item 7 only
[
make object! [id: "aa" items: [5] ]
]
THE SMALL PRINT
** You can assume the data block always contains at least one object
** You can also assume that there is at least one data item (ie a
least one of the objects will have at least one entry in its
'items block)
** you can assume that the range block has been correctly coerced
to match the data, so:
[51 74] -- means there are at least 74 items in the data
[74 51] -- won't happen: the second number will never
be lower than the first
** the objects must emerge in the same sequence as they began
** (the above three assumptions means that there will always be at
least one object with at least one item in the output)
** the items must emerge in the same sequence as they began
** do not emit any objects if they have zero selected items
** there may be several thousand entries in each 'items block, so
any solution that runs a crude loop will be very slow
** you can directly edit the original block and objects, or create
your own copies
THE JUDGES
None, really -- if you contribute a solution, feel free to comment
on other people's solutions.
THE PRIZE!
There is no prize other than the glory of having written and
published an elegant solution.
I would like permission to use the best R2 solution on REBOL.org
(its part of a search CGI) to replace klunky the code I wrote myself.
If you enter .... Good luck!
Sunanda
--
To unsubscribe from the list, just send an email to
lists at rebol.com with unsubscribe as the subject.
