How to use boost::preprocessor to unzip a sequence?
Asked Answered
N

2

5

How to use boost::preprocessor to unzip a sequence of pairs?

For example, I have a sequence as below (comma between doesn't matter)

(int,x)(double,y)(float,z) or
(int,x),(double,y),(float,z) or
((int)(x))((double)(y))((float)(z))

and want to convert to

int,double,float

and

x,y,z

By using macor like

UNZIP(i, seq) 

where i is the index.

Nikkinikkie answered 20/10, 2014 at 21:44 Comment(1)
You have a comma between the last two elements and no comma between the first two. Which do you prefer? Actually, it's looking a lot better with the commas.Bollworm
L
7

Unzipping of (int, x, 10)(double, y, 20)(float, z, 30), i.e. sequence without commas between elements.

LIVE DEMO

#include <boost/preprocessor/punctuation/comma_if.hpp>
#include <boost/preprocessor/seq/for_each_i.hpp>
#include <boost/preprocessor/seq/pop_front.hpp>
#include <boost/preprocessor/seq/for_each.hpp>
#include <boost/preprocessor/variadic/elem.hpp>
#include <boost/preprocessor/cat.hpp>

// Such technique is used at:
//   http://www.boost.org/doc/libs/1_56_0/boost/fusion/adapted/struct/define_struct.hpp
#define AUXILIARY_0(...) ((__VA_ARGS__)) AUXILIARY_1
#define AUXILIARY_1(...) ((__VA_ARGS__)) AUXILIARY_0
#define AUXILIARY_0_END
#define AUXILIARY_1_END

#define REMOVE_PARENTHESES(...) __VA_ARGS__

#define COMMA_SEPARATED(r, data, i, elem) \
    BOOST_PP_COMMA_IF(i) BOOST_PP_VARIADIC_ELEM(data, REMOVE_PARENTHESES elem) \
/**/

#define ZIPPED_TO_SEQ(zipped) \
    BOOST_PP_SEQ_POP_FRONT(BOOST_PP_CAT(AUXILIARY_0(0)zipped,_END)) \
/**/

#define FOR_EACH_ZIPPED_I(macro, data, zipped) \
    BOOST_PP_SEQ_FOR_EACH_I(macro, data, ZIPPED_TO_SEQ(zipped)) \
/**/

#define UNZIP(i, zipped) FOR_EACH_ZIPPED_I(COMMA_SEPARATED, i, zipped)

/*******************************************************************/
// DEMO:

#define zipped (int, x, 10)(double, y, 20)(float, z, 30)

FIRST:  UNZIP(0, zipped)
SECOND: UNZIP(1, zipped)
THIRD:  UNZIP(2, zipped)

Preprocessor output:

FIRST: int , double , float
SECOND: x , y , z
THIRD: 10 , 20 , 30
Laurice answered 20/10, 2014 at 22:10 Comment(0)
B
5

Assuming you meant commas ((int,x),(double,y),(float,z)), Boost.PP works great:

#include <boost/preprocessor.hpp> //or smaller individual headers

//SEQ_ENUM adds commas between each element of the sequence
//the sequence is transformed from what's passed into the ...
//the invocation of UNZIP_MACRO is given the index to use in each element as data
#define UNZIP(i, ...)                             \
    BOOST_PP_SEQ_ENUM(                            \
        BOOST_PP_SEQ_TRANSFORM(                   \
            UNZIP_MACRO,                          \
            i,                                    \
            BOOST_PP_VARIADIC_TO_SEQ(__VA_ARGS__) \
        )                                         \
    )

//called with each element of the sequence
#define UNZIP_MACRO(s, data, elem) \
    BOOST_PP_TUPLE_ELEM(data, elem)


#define ZIPPED_SEQUENCE (int,x),(double,y),(float,z)

UNZIP(0, ZIPPED_SEQUENCE) //int, double, float
UNZIP(1, ZIPPED_SEQUENCE) //x, y, z
Bollworm answered 20/10, 2014 at 22:8 Comment(0)

© 2022 - 2024 — McMap. All rights reserved.