Perhaps: rad=: (#~ 1 - ] -:"1 |.!.__~&1)&.|:
Note that I have assumed that you will not be using negative infinity in B. -- Raul On Thu, Oct 4, 2012 at 8:12 PM, Linda Alvord <lindaalv...@verizon.net> wrote: > I think I only need "ifany" . But I could call it "ifempty" . > > ifany=: ^:(*@#@]) > rad=: 13 :'(1,-. 2-:/\|:y)#"1 y' > ]B=:2 4 $3 5 5 4 2 1 1 6 > 3 5 5 4 > 2 1 1 6 > rad ifany B > 3 5 4 > 2 1 6 > rad ifany C=:'' > > Linda > > -----Original Message----- > From: programming-boun...@forums.jsoftware.com > [mailto:programming-boun...@forums.jsoftware.com] On Behalf Of Henry Rich > Sent: Thursday, October 04, 2012 7:38 PM > To: programm...@jsoftware.com > Subject: Re: [Jprogramming] Why the length error? > > I use > > ifany=: ^:(*@#@]) > butifnull =: 2 : 'v"_`u@.(*@#@])' > > rad ifany C > > rad butifnull (2 0$0) C > > Henry Rich > > On 10/4/2012 7:22 PM, Linda Alvord wrote: >> I just realized that all these choices, change B in some way. I want >> it to always be a two row table of pairs in columns. >> >> B=:2 4 $3 5 5 4 2 1 1 6 >> 3 5 5 4 >> 2 1 1 6 >> >> ''$ B >> 3 5 5 4 >> {. B >> 3 5 5 4 >> 0 { B >> 3 5 5 4 >> ,/ B >> 3 5 5 4 2 1 1 6 >> >> My only concern is when B happens to have no pairs, in which case I >> want a result that is an empty array when rad is used. >> >> I'm glad I noticed these problems as they would have caused even more > grief. >> >> Thanks for you continuing assistance. >> >> Linda >> >> -----Original Message----- >> From: programming-boun...@forums.jsoftware.com >> [mailto:programming-boun...@forums.jsoftware.com] On Behalf Of Raul >> Miller >> Sent: Thursday, October 04, 2012 11:34 AM >> To: programm...@jsoftware.com >> Subject: Re: [Jprogramming] Why the length error? >> >> Here are some options (I'm repeating concepts expressed by Roger Hui >> and Devon McCormick: >> >> L=: 1,'' >> ''$ L >> {. L >> 0 { L >> ,/ L >> >> Roger's suggestion is equivalent to the APL classic in this context, >> but in the general case you would need to ravel the right argument if >> L had a rank greater than 1 (APL's rho is like ($,) in J). >> >> {. L and ''$L and 0 { L do exactly the same thing, but the syntax is >> different, which can be significant if you are building a train >> (Roger's approach re-expresses nicely as a fork -- 0 { L is also nice >> in a fork but since numbers stick to other numbers to form numeric >> words it's not quite as nice as Roger's approach).. >> >> ,/ L is different -- the others will discard extra elements and always >> reduce the rank of the result while ,/ removes a leading dimension but >> never gets rid of any elements (so if you have multiple elements it >> will not take you below rank 1). >> >> Anyways... simplest depends on you and where you are going. And note >> that there's other options also, depending... (+/L for example, would >> also get you an atomic 1 in this case).. >> >> -- >> Raul >> >> On Thu, Oct 4, 2012 at 9:35 AM, Linda Alvord <lindaalv...@verizon.net> >> wrote: >>> This is the crux of the problem: >>> >>> 1,'' >>> 1 >>> $1 >>> $1,'' >>> 1 >>> >>> What is the simplest way to make the shape of 1,'' be empty? >>> >>> Linda >>> >>> >>> -----Original Messapege----- >>> From: programming-boun...@forums.jsoftware.com >>> [mailto:programming-boun...@forums.jsoftware.com] On Behalf Of Ric >>> Sherlock >>> Sent: Thursday, October 04, 2012 2:40 AM >>> To: programm...@jsoftware.com >>> Subject: Re: [Jprogramming] Why the length error? >>> >>> This gives a length error because >>> (1,'') # '' >>> |length error >>> | (1,'') #'' >>> >>> An simple solution to your problem is to returns the input on error: >>> rad=: (#~ [: -. 0 , 2 -:/\ ])@|: :: ] >>> rad '' >>> >>> rad B >>> 3 2 >>> 5 1 >>> 4 6 >>> >>> On Thu, Oct 4, 2012 at 6:41 PM, Linda Alvord >>> <lindaalv...@verizon.net> >>> wrote: >>>> >>>> rad=: 13 :'(A=:1,-. 2-:/\|:y)#"1 y' NB. Remove adjacent >>>> duplicates >>>> >>>> ]B=:2 4 $3 5 5 4 2 1 1 6 >>>> 3 5 5 4 >>>> 2 1 1 6 >>>> rad B >>>> 3 5 4 >>>> 2 1 6 >>>> >>>> >>>> The verb rad works as expected, but suppose in an earlier step B >>>> has become empty, I get a length error. Why does this happen? >>>> >>>> >>>> C=:'' >>>> rad C >>>> |length error: rad >>>> | (A=:1,-.2-:/\|:y) #"1 y >>>> A >>>> 1 >>>> 1#"1 C >>>> >>>> What can I do to get an empty result? >>>> >>>> Linda >>>> >>>> -------------------------------------------------------------------- >>>> - >>>> - 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 >> ---------------------------------------------------------------------- >> 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 > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
