Of course, you can still go by the dot product. Just don't sum: a (] . *) |:b 10 14 22 30 36 48 52 68
On Mon, Apr 18, 2016 at 11:49 AM, EelVex <eel...@gmail.com> wrote: > If you remove the first axis of a you can multiply-1: > > (,a) *"1 b > 10 22 36 52 > 14 30 48 68 > > Do you really need a to be 1xn ? > > > On Mon, Apr 18, 2016 at 11:41 AM, 'Jon Hough' via Programming < > programm...@jsoftware.com> wrote: > >> I am sure there is a simple answer to this, but I would like to know it. >> >> Say I have two matrices: >> >> a =: 1 4 $ 1 2 3 4 >> b =: 2 4 $ 10 11 12 13 14 15 16 17 >> >> and I want to multiply a's single row element-wise with each of b's rows. >> Not matrix multiplication, so the resulting matrix should have dimensions 2 >> 4, same as b. >> >> Obviously a * b gives a length error, >> >> So does a *"1 1 b >> >> My best solution is pretty ugly, >> >> (($b)$,a ) * b >> >> i.e. reshape a so it has the same number of rows as b, which duplicates >> the items of its top row into the second row and then multiply row-wise. >> >> Is there a better way? >> Incidentally, isn't a*1 1 b a row-to-row multiplication. I thought this >> would have worked, and am slightly confused why it doesn't. >> ---------------------------------------------------------------------- >> For information about J forums see http://www.jsoftware.com/forums.htm > > > ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
