www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - Switching rows with columns

reply "Tanel =?UTF-8?B?VGFnYXbDpGxpIg==?= <tanel58 hotmail.com> writes:
I have a range of ranges and need to change it so the elements 
are column-aligned instead of row-aligned.

For example,
[[1, 2, 3],
  [4, 5, 6],
  [7, 8, 9]]
would change into
[[1, 4, 7],
  [2, 5, 8],
  [3, 6, 0]].

Can I even do this with ranges, and if so, how?
Jul 04 2015
parent reply "Marc =?UTF-8?B?U2Now7x0eiI=?= <schuetzm gmx.net> writes:
On Saturday, 4 July 2015 at 15:28:56 UTC, Tanel Tagaväli wrote:

 I have a range of ranges and need to change it so the elements 
 are column-aligned instead of row-aligned.

 For example,
 [[1, 2, 3],
  [4, 5, 6],
  [7, 8, 9]]
 would change into
 [[1, 4, 7],
  [2, 5, 8],
  [3, 6, 0]].

 Can I even do this with ranges, and if so, how?


Try std.range.transposed:
Jul 04 2015
parent reply "Tanel =?UTF-8?B?VGFnYXbDpGxpIg==?= <tanel58 hotmail.com> writes:
On Saturday, 4 July 2015 at 16:29:44 UTC, Marc Schütz wrote:

 Try std.range.transposed:



Does this work with user-defined types?

I defined two structs that implement the InputRange(possibly 
ForwardRange) interface, an integer range and a range of those 
integer ranges.

DMD tells me the "transposed" template cannot deduce function 
from the arguments, which is a range of integer ranges.
The template does, however, work on a two-dimensional array.

The "save" method is implemented using a simple "return this", 
could this be to blame?
Jul 04 2015
parent reply "Marc =?UTF-8?B?U2Now7x0eiI=?= <schuetzm gmx.net> writes:
On Sunday, 5 July 2015 at 00:18:18 UTC, Tanel Tagaväli wrote:

 On Saturday, 4 July 2015 at 16:29:44 UTC, Marc Schütz wrote:

 Try std.range.transposed:



 Does this work with user-defined types?

 I defined two structs that implement the InputRange(possibly 
 ForwardRange) interface, an integer range and a range of those 
 integer ranges.

 DMD tells me the "transposed" template cannot deduce function 
 from the arguments, which is a range of integer ranges.
 The template does, however, work on a two-dimensional array.

 The "save" method is implemented using a simple "return this", 
 could this be to blame?


This should be fine. Which of the ranges is the ForwardRange? It 
should be the range-of-ranges.

You can check which of the template constraints fails:

import std.range;
alias RangeOfRanges = typeof(myRangeOfRanges);
pragma(msg, isForwardRange!RangeOfRanges);
pragma(msg, isInputRange!(ElementType!RangeOfRanges));
pragma(msg, hasAssignableElements!RangeOfRanges);

Maybe it's the last condition, `hasAssignableElements`. I don't 
know whether that one is really necessary...
Jul 05 2015
parent "Tanel =?UTF-8?B?VGFnYXbDpGxpIg==?= <tanel58 hotmail.com> writes:
On Sunday, 5 July 2015 at 11:35:14 UTC, Marc Schütz wrote:

 Maybe it's the last condition, `hasAssignableElements`. I don't 
 know whether that one is really necessary...


It probably is. The last check returned false.
Since I already would have to implement "r.front = ", I'll just 
use arrays.

Thanks for the help.
Jul 05 2015