mpl: generate kind of permutations
Hello,
I have 2 strictly ordered sequences of dates t1... tn, and td1...tdm, where
m and n may or may not be different.
We always have: t1<...
Hicham Mouline wrote:
I have 2 strictly ordered sequences of dates t1... tn, and td1...tdm, where m and n may or may not be different. We always have: t1<...
I am trying to place the (t1...tn) sequence over the (td1...tdm) sequence, and generate all these "permutations" ie
<snip examples>
I hope the description is clear enough even though it is not formal at all.
I'm sorry, maybe I wasn't trying hard enough to understand what you're trying to build here. I'm left with a hazy impression that you want to construct a merged sequence of dates, possibly including tag information identifying the original sequence from which a particular date came. But my notion of merge processing includes the idea of "consuming" items from the input sequences. I don't know how I'd attempt to implement that at compile time. Then again, I'm the wrong person to ask about MPL. Essentially I'm responding that no, I really didn't understand your problem statement. Perhaps a clearer restatement would prompt useful suggestions from others?
-----Original Message----- From: boost-users-bounces@lists.boost.org [mailto:boost-users- bounces@lists.boost.org] On Behalf Of Nat Goodspeed Sent: 05 January 2010 3:47 To: boost-users@lists.boost.org Subject: Re: [Boost-users] mpl: generate kind of permutations
Hicham Mouline wrote:
I have 2 strictly ordered sequences of dates t1... tn, and td1...tdm, where m and n may or may not be different. We always have: t1<...
I am trying to place the (t1...tn) sequence over the (td1...tdm) sequence, and generate all these "permutations" ie
<snip examples>
I hope the description is clear enough even though it is not formal at all.
I'm sorry, maybe I wasn't trying hard enough to understand what you're trying to build here. I'm left with a hazy impression that you want to construct a merged sequence of dates, possibly including tag information identifying the original sequence from which a particular date came.
But my notion of merge processing includes the idea of "consuming" items from the input sequences. I don't know how I'd attempt to implement that at compile time. Then again, I'm the wrong person to ask about MPL.
Essentially I'm responding that no, I really didn't understand your problem statement. Perhaps a clearer restatement would prompt useful suggestions from others? Taking my example from before, starting the dates vector with YYYYMMDD int elements mpl::vector_c< 20100109, 20100125, 20100322, 20100428 > tds; // requires strictly sorted elements
and I want to place n=3 dates in buckets formed by the above 4 dates. Let's say the way I choose the dates is 1 day, then 2 day, then 3days, less the dates in tds. So I want the metafunction to return the list of all mpl::vector_c<> that have always 3 elements strictly sorted, so that . all 3 dates < 20100109 ( 20100106 20100107 20100108 ) . 2 dates < 20100109 and 3rd date=20100109 ( 20100107 20100108 20100109 ) . ( 20100107 20100108 20100110 ) . ( 20100107 20100108 20100125 ) . ( 20100107 20100108 20100126 ) . ( 20100107 20100108 20100322 ) . ( 20100107 20100108 20100323 ) . ( 20100107 20100108 20100428 ) . ( 20100107 20100108 20100429 ) . 1 date < 20100109 and 2dates>=20100109 ( 20100108 20100109 20100110 ) . ( 20100108 20100109 20100125 ) . ( 20100108 20100109 20100126 ) . ( 20100108 20100109 20100322 ) . ( 20100108 20100109 20100323 ) . ( 20100108 20100109 20100428 ) . ( 20100108 20100109 20100429 ) . ( 20100108 20100110 20100111 ) . ( 20100108 20100110 20100125) . ( 20100108 20100110 20100126 ) . ( 20100108 20100110 20100322 ) . ( 20100108 20100110 20100323 ) . ( 20100108 20100110 20100428 ) . ( 20100108 20100110 20100429 ) Etc Etc 0 date<20100109 and 3dates>=20100109 The objective is to write the metafunction for generic tds vector, and generic n (not just 3) I hope this is clearer otherwise, I will write the complete list next time. Regards,
AMDG Hicham Mouline wrote:
Taking my example from before, starting the dates vector with YYYYMMDD int elements mpl::vector_c< 20100109, 20100125, 20100322, 20100428 > tds; // requires strictly sorted elements
and I want to place n=3 dates in buckets formed by the above 4 dates.
Let's say the way I choose the dates is 1 day, then 2 day, then 3days, less the dates in tds.
So I want the metafunction to return the list of all mpl::vector_c<> that have always 3 elements strictly sorted, so that
. all 3 dates < 20100109 ( 20100106 20100107 20100108 )
. 2 dates < 20100109 and 3rd date=20100109 ( 20100107 20100108 20100109 ) . ( 20100107 20100108 20100110 ) . ( 20100107 20100108 20100125 ) . ( 20100107 20100108 20100126 ) . ( 20100107 20100108 20100322 ) . ( 20100107 20100108 20100323 ) . ( 20100107 20100108 20100428 ) . ( 20100107 20100108 20100429 )
. 1 date < 20100109 and 2dates>=20100109 ( 20100108 20100109 20100110 ) . ( 20100108 20100109 20100125 ) . ( 20100108 20100109 20100126 ) . ( 20100108 20100109 20100322 ) . ( 20100108 20100109 20100323 ) . ( 20100108 20100109 20100428 ) . ( 20100108 20100109 20100429 )
. ( 20100108 20100110 20100111 ) . ( 20100108 20100110 20100125) . ( 20100108 20100110 20100126 ) . ( 20100108 20100110 20100322 ) . ( 20100108 20100110 20100323 ) . ( 20100108 20100110 20100428 ) . ( 20100108 20100110 20100429 ) Etc Etc
0 date<20100109 and 3dates>=20100109
The objective is to write the metafunction for generic tds vector, and generic n (not just 3)
I hope this is clearer otherwise, I will write the complete list next time.
Something like this should generate all subsets of a
sequence of a specific size. (Warning completely untested)
template
participants (3)
-
Hicham Mouline -
Nat Goodspeed -
Steven Watanabe