I would start by better understanding the specific problem I was trying to solve.
(As opposed to coming up with a solution and then looking for a problem that fits to it...) Thanks, -- Raul On Thu, Sep 21, 2017 at 9:35 AM, Erling Hellenäs <[email protected]> wrote: > Hi all! > > I made a mistake when trying to get an error obviously caused by a platform > exception. > Databases are full of fields without data. Doesn't R handle this in a good > way? > What is your proposed way of handling this missing data? > You consider NaNs, positive and negative infinity as unknown values too? > Would you prefer an error instead of getting these? Or you just look for > them in the tests and make sure they never happen in production? > > /Erling > > > Den 2017-09-21 kl. 14:41, skrev Don Guinn: >> >> Did you mean? >> >> 1/0 >> |domain error >> | 1 /0 >> |[-0] >> 1%0 >> _ >> >> What made me think of sparse arrays is that it does much of the leg work >> to >> identify some number, like zero, to remove. It keeps an index list of >> elements not removed and has tools to extract only those elements not >> removed. It's really not what sparse ($.) is intended for, but I thought >> it >> might be useful in this case. >> >> In my thinking, unknown values should never be put in an array. Having >> them >> in an array leads to so many problems, only a few have been mentioned in >> this thread. >> >> If "unknown values" are generated as part of the computation then it seems >> to me it is time to go back and look at the problem analysis. >> ---------------------------------------------------------------------- >> For information about J forums see http://www.jsoftware.com/forums.htm > > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
