from vector to matrix

[Jay Rege]

Hey,
I want to transform the vector x into a Matrix A with 4 columns.
I want to do that since I want to build a function, which sums up every fourthed element of the vector.I guess if I would create this matrix I could just use colsum(x) and than display col4.
Any help would be great!

[Niels Henrik Bruun]

Hi Jay
Two alternatives. The first handles if you only need every fourth value, the second is getting the matrix you asked for (but may not be needing):

: x = round(runiform(100,1) * 100)    // Generating a vector

: select(x, !mod(1::rows(x), 4))    // Selecting every fourth value
         1
     +------+
   1 |  59  |
   2 |  10  |
   3 |  83  |
   4 |  53  |
   5 |  42  |
   6 |  52  |
   7 |   9  |
   8 |   4  |
   9 |  45  |
  10 |  71  |
  11 |  60  |
  12 |  75  |
  13 |  37  |
  14 |  42  |
  15 |  59  |
  16 |  75  |
  17 |  19  |
  18 |  78  |
  19 |  92  |
  20 |  63  |
  21 |  18  |
  22 |  17  |
  23 |  26  |
  24 |  92  |
  25 |  87  |
     +------+

: colshape(x, 4)    // Creating the matrix you asked for
         1    2    3    4
     +---------------------+
   1 |  69   98   67   59  |
   2 |  80   78   65   10  |
   3 |  69   87   53   83  |
   4 |  93   17   55   53  |
   5 |  78   13   28   42  |
   6 |  14   33   47   52  |
   7 |   7    7   68    9  |
   8 |  13   87   25    4  |
   9 |  38   77   76   45  |
  10 |  41   30   68   71  |
  11 |  57   18   11   60  |
  12 |  63   63   99   75  |
  13 |  17   61   58   37  |
  14 |  30    1   67   42  |
  15 |  95    9   89   59  |
  16 |  40   67   42   75  |
  17 |  72   85   79   19  |
  18 |  39   24   34   78  |
  19 |  75   23   17   92  |
  20 |  31   90    8   63  |
  21 |  81   88    3   18  |
  22 |  58   41   62   17  |
  23 |  36   13    0   26  |
  24 |  65   93   82   92  |
  25 |  75   52   40   87  |
     +---------------------+

: end

[Jay Rege]

Thanks a lot!
Maybe one further question.
Do you know how to sum these elements up?
I`ve tried sum(select(x, !mod(1::rows(x), 4))) but that returns 0
furthermore I´ve tried to store the selected result and than sum it up,
x = (1, 3, 4, 4, 5, 6)
y = select(x, !mod(1::rows(x), 4))
sum(y)
which did not work as well.
Thanks again!

[Niels Henrik Bruun]

Your x is a row vector. The function select needs a column vector, so you have to transpose.
Also for a row vector you need to count columns (mistakes I also do all the time ).
Look at the code:

: x = 1..50

: y = select(x', !mod(1::cols(x), 4))

: y
         1
     +------+
   1 |   4  |
   2 |   8  |
   3 |  12  |
   4 |  16  |
   5 |  20  |
   6 |  24  |
   7 |  28  |
   8 |  32  |
   9 |  36  |
  10 |  40  |
  11 |  44  |
  12 |  48  |
     +------+

: sum(y)
  312

Now it works

[Jay Rege]

Nice! Thanks again!
Is it possible to make it irrelevant wether its a volume or row vector?

[Niels Henrik Bruun]

I don't think so. Some sort of consistency is required