On Thu, Aug 7, 2008 at 3:06 PM, Roger Hui <[EMAIL PROTECTED]> wrote:
> [EMAIL PROTECTED] doesn't sound right to me. {.x E. y is 1 iff x is a prefix
> of y, and for that computation x -: (#x){.y is much more efficient.
> Perhaps Raul meant to say something like (E.i.1:), the index
> of the first occurrence. This latter phrase happens to be
> supported by special code.
I do not know enough about the width of the data, nor
the implementation of E. to know if ([ -: [EMAIL PROTECTED] {. ])"1 would
be more efficient than [EMAIL PROTECTED]"1 and I have not thought about
phrases like (] #~ [: *./@:-.@(0&=)@:> (E.i.1:)"1 L:0)
[But the requirement was to reject rows of a character
table when they were prefixed with one of a set of
specific character sequences, so {.x E. y does achieve
the desired end.]
--
Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
