I checked ... the "each" operator works exactly the same in A+ with a defined function: The code in the defined function is re-interpreted for each element.
I had used the "each" operator primarily with primitive functions, and so, never noticed the load it could put on the interpreter when used with a defined function on an entity containing a large number of elements. On Sun, Jan 8, 2012 at 8:36 PM, Tom Szczesny <tav...@gmail.com> wrote: > Yes -- > the conjuction explanation makes a lot more sense ... > The application of cutl a 'record' at a time becomes apparent before the > first line of code in cutl is executed. > > load 'debug' > dbss 'cutl *:*' > dbr 1 > ','&cutl;._2 csvfile > |stop: cutl > | y=.y,x > |cutl[:0] > y > a,tkr,dt,s,td,l,cd,qu,pr,val,cst,ty,lgl,sgl,iq,iu,ix,qc,qp,st,dte,cq,qv,cmt > x > , > > From A+, I'm familiar with a function being applied to each element of an > entity. > What surprised me was that the code in cutl was being reinterpreted by the > J-interpreter for each element. > I guess that's because the the conjuction is applied to a defined > function, and not a primitive. > > > On Sun, Jan 8, 2012 at 5:06 PM, Ric Sherlock <tikk...@gmail.com> wrote: > >> To be precise cutl is being "fed" csvfile one record at a time by the >> conjunction cut ( ;. ). ( cutl;._2 ) will chop its right argument into >> bits using the last item of the right argument (probably a line-feed) >> as the record delimiter (fret) and apply the verb u (in your case ( >> ','&cutl ) to each of those records. >> >> http://www.jsoftware.com/docs/help701/dictionary/d331.htm >> >> On Mon, Jan 9, 2012 at 6:50 AM, Tom Szczesny <tav...@gmail.com> wrote: >> > >> > ',' &cutl;._2 csvfile >> > >> > I was amazed that cutl loops through csvfile one 'record' at a time, >> > as none of the 9 control structure patterns listed in the Primer were >> > evident. >> > >> > >> > >> > On Mon, Jan 2, 2012 at 12:44 PM, Joey K Tuttle <j...@qued.com> wrote: >> > >> >> I believe that there are a collection of csv related utilities in >> addons >> >> ( require 'csv' will make those available to you). >> >> >> >> However, for years I have used a simplistic verb to bring .csv files >> >> into a j session as a table of enclosed values - >> >> >> >> cutl =: 3 : 0 >> >> : >> >> y =. y,x >> >> q =. y e.'"' >> >> qs =. ~:/\q >> >> }.^: ('"'&=@([:{.1:{.]))&.> (qs<y e.x) <;._2&((q<:qs)&#) y >> >> ) >> >> >> >> tablecsv =: 3 : 0 >> >> ',' tablecsv y NB. allows choosing an alternative delimiter >> >> : >> >> x &cutl;._2 y >> >> ) >> >> >> >> >> >> On 2012/01/02 09:22 , Tom Szczesny wrote: >> >> > Some languages, like q, have a built-in command for reading, >> parsing, and >> >> > loading data from csv files: >> >> > t: ("SSDIDSSIFFFIS";enlist ",") 0: `:/home/tom/t.csv >> >> > >> >> > Other languages, like a+, do not, and the user needs to write a >> function >> >> to >> >> > read, parse and load data from csv files. >> >> > >> >> > > From a cursory review, it appears that j is in the a+ camp. >> >> > Please confirm that such a facility does not exist in j, before I >> start >> >> > writing my own. >> >> > >> >> >> >> ---------------------------------------------------------------------- >> >> 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
