Also, what's up with this:
lower =: ]`(a.{~32&+@:(a.&i.)) @. (>&64 *. <&91)@:(a.&i.) "0
lower 'A'
|index error: lower
| lower'A'
0:`1: @. (>&64 *. <&91)@:(a.&i.) 'A'
1
]`(a.{~32&+@:(a.&i.)) @. 1: 'A'
a
]`(a.{~32&+@:(a.&i.)) @. (>&64 *. <&91)@:(a.&i.) 'A'
|index error
| ]`(a.{~32&+@:(a.&i.))@.(>&64*.<&91)@:(a.&i.)'A'
The verb on the right of agenda produces 1 for 'A' on its own, and the second
gerund produces 'a' from 'A' on its own, but when combined I get an index error.
On 28 Oct 2011, at 16:21, David Vaughan wrote:
> I was trying to write a verb to 'clean' inputs by removing any non-alphabetic
> characters and converting to lowercase. What I came up with seems to work,
> however I'm getting issues with spelling errors again, due to (I assume)
> invisible unicode characters.
>
> This is copied and pasted from my .ijs file:
>
> tolower =: 3 :'if. (>&64 *. <&91) a.i.y do. a. {~ 32 + a.i.y else. y end.'"0
> string_clean =: tolower #~ ([: (>&96 *. <&123) a.&i.@:tolower)
> string_clean 'TeSt.!'
>
> It works if I load it and then run string_clean, but not if I try to use 0!:1:
>
> 0!:1<'J/scripts/charstuff.ijs'
> tolower =: 3 :'if. (>&64 *. <&91) a.i.y do. a. {~ 32 + a.i.y else. y
> end.'"0
> |spelling error
> |
> | ^
> | tolower=: 3 :'if. (>&64 *. <&91) a.i.y do. a. {~ 32 + a.i.y else. y
> end.'"0
> |[-0] J/scripts/charstuff.ijs
>
> I had a fix for this sort of spelling error that seems to work whenever I use
> load thanks to Raul Miller:
>
> script_z_
> [: 3 : '0!:0 -.&(128}.a.) 1!:1 y [ 4!:55<''y''' jpath_z_&.:>
>
> That is in profile.ijs.
>
> Can anyone help me?
>
> Thanks.
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
