Re: [Jprogramming] converting numbers and rank trouble

Nice - I didn't realize dyalog had a blog. Looks like great content. It may be time to finally learn a bit of APL to have a passing reading comprehension in it. I have stopped and started the Dyalog book a few times over the years as I have wanted to digest some of the Essays but that is probably b

Re: [Jprogramming] converting numbers and rank trouble

Sorry for the confusion. Yes, it is a character matrix. In this case it is a text file with one line per numeric value that is read with fread. I don't box the values I was using boxed arrays as a shorthand to create the character matrix sample. I find it easier sometimes to do that for sample - e

Re: [Jprogramming] converting numbers and rank trouble

Not nested boxes (for example). But point taken. Thanks, -- Raul On Fri, Apr 13, 2018 at 6:11 PM, Roger Hui wrote: > They do _under_ conversion to numbers and under open. > > > > On Fri, Apr 13, 2018 at 2:49 PM, Raul Miller wrote: > >> I remember at one point wishing that these would all wo

Re: [Jprogramming] converting numbers and rank trouble

They do _under_ conversion to numbers and under open. On Fri, Apr 13, 2018 at 2:49 PM, Raul Miller wrote: > I remember at one point wishing that these would all work on character > and boxed data. > > Thanks, > > -- > Raul > > > On Fri, Apr 13, 2018 at 4:25 PM, Roger Hui > wrote: > > It's use

Re: [Jprogramming] converting numbers and rank trouble

I remember at one point wishing that these would all work on character and boxed data. Thanks, -- Raul On Fri, Apr 13, 2018 at 4:25 PM, Roger Hui wrote: > It's useful to have at your fingertips a (small) set of manoeuvers: > > x * 0 = 0 > x * 1 = x > > x ^ 0 = 1 > x ^ 1 = x > > etc. > > The

Re: [Jprogramming] converting numbers and rank trouble

It's useful to have at your fingertips a (small) set of manoeuvers: x * 0 = 0 x * 1 = x x ^ 0 = 1 x ^ 1 = x etc. The Y {~ X i. x phrase, for mapping cells in X to values in Y, is also generally useful. See the Dyalog APL blog post Calcultion v Look-Up

Re: [Jprogramming] converting numbers and rank trouble

I had trouble myself understanding what Joe was driving at, and that lead to my mistaken call for better sample data. What Joe is working with is a character matrix - one row of characters represents one number. So there will be multiple columns in the source data and that corresponds to a list of

Re: [Jprogramming] converting numbers and rank trouble

Roger said that the boxed sample was only for illustration. But then, where does the empty box come from? Or is this a comma or tab delimited file? Granted, if the original file is pretty trashy then it is probably necessary to do some clean-up ending out in table of boxes. But if the data is relat

Re: [Jprogramming] converting numbers and rank trouble

Don - the source data is a text file of float data that is being read into J Roger - thanks, for the confirmation. I find that when I'm way from J/APL for awhile, my mind drifts back into "if/then" thinking instead of the simpler and faster method of using multiplication where possible. I fiddled

Re: [Jprogramming] converting numbers and rank trouble

Yes, it's a boon that 0 * x is 0 for any x. Some authors go even further: Knuth propose to have 0 * x be 0 even when x is an expression which is in error. He wanted this to make "Iverson's convention" work in more situations. See https://arxiv.org/PS_cache/math/pdf/9205/9205211v1.pdf On Fri,

Re: [Jprogramming] converting numbers and rank trouble

He's using boxed arrays only to generate the small sample data. On Fri, Apr 13, 2018 at 12:43 PM, Don Guinn wrote: > Just out of curiosity, surely the data didn't start out as an array of > boxes. Perhaps it might be faster not to have everything boxed. Validation > could be done once on the or

Re: [Jprogramming] converting numbers and rank trouble

Just out of curiosity, surely the data didn't start out as an array of boxes. Perhaps it might be faster not to have everything boxed. Validation could be done once on the original source. On Fri, Apr 13, 2018, 1:24 PM Joe Bogner wrote: > Well, I stumbled upon a fast way that takes advantage of

Re: [Jprogramming] converting numbers and rank trouble

There is no faster way, although you can speed up "isblank": [: -. {.@(0&{.) -:"1 ] It's wasteful to catenate ' ' to a great big array only to take and drop its first cell. On Fri, Apr 13, 2018 at 12:24 PM, Joe Bogner wrote: > Well, I stumbled upon a fast way that takes advantage of simple

Re: [Jprogramming] converting numbers and rank trouble

Well, I stumbled upon a fast way that takes advantage of simple multiplication _ * 0 = 0 _ * 1 = _ 5 * 1 = 5 _3 {. (-.@isblank * _&".) c1 1.01 _ 0 (6!:2) '(isblank * _&".) c1' 1.00593 I'm still interested in alternatives that are similar speed if possible On Fri, Apr 13, 2018 at 3:19

Re: [Jprogramming] converting numbers and rank trouble

Thank you -- Yes, I didn't want to send along the full sample, although I could have created some dummy data. Both Don and Raul's method are similar and are equally slow as the way I had It's not terribly important but I have like 10 numeric columns that I convert in a script and have to painfull

Re: [Jprogramming] converting numbers and rank trouble

Your sample data does not match your problem description. That said, {.@(_&".)"1 > ' ';'bad';'10' 0 _ 10 Thanks, -- Raul On Fri, Apr 13, 2018 at 2:03 PM, Joe Bogner wrote: > I have a large byte array that I want to convert to a number array. I want > to use _ to indicate a bad value, but

Re: [Jprogramming] converting numbers and rank trouble

Is this what you want? _".&>'';'bad';'10' 0 _ 10 On Fri, Apr 13, 2018 at 12:03 PM, Joe Bogner wrote: > I have a large byte array that I want to convert to a number array. I want > to use _ to indicate a bad value, but blank should be treated as 0 > > $ c > 4862343 10 > > Instead of this: