don't know if
unrolling a tree is a formal term, but it ends up being flat at the end of the
process.
- Original Message -
From: Raul Miller
To: Programming forum
Cc:
Sent: Thursday, September 4, 2014 12:14:22 AM
Subject: Re: [Jprogramming] A symetrical hash function that outputs va
On Wed, Sep 3, 2014 at 10:28 PM, 'Pascal Jasmin' via Programming
wrote:
> The difference between using L:1 and S:1 is that with L:1 the tree
> structure is kept. When used recursively, the tree gets rolled up
> from the bottom. If there is a relationship with the group of cells
> at a tree level
t the most important part was the any
shape. Keys, even compound keys, tend not to get that long.
- Original Message -
From: bill lam
To: 'Pascal Jasmin' via Programming
Cc:
Sent: Wednesday, September 3, 2014 10:56:34 PM
Subject: Re: [Jprogramming] A symetrical hash function
J implementation places an artificial limit on the lengths of
names. try
(255$'a')=.0
(256$'a')=.0
|limit error
| (256$'a')=.0
Ср, 03 сен 2014, jprogramming написал(а):
> strbracket =: {.@:[ , ] , {:@[
> joinB =: (1 : ' ((- # m)&}.@;@(,&m&.>"1))')(@: (] : ;))
> nums2name =: 'NUM' joi
sage -
From: Raul Miller
To: Programming forum
Cc:
Sent: Wednesday, September 3, 2014 9:07:42 PM
Subject: Re: [Jprogramming] A symetrical hash function that outputs valid J
variable names
Hmm...
A few comments:
1] I like strbracket:
'()' strbracket 'abc'
(abc)
2] If
Hmm...
A few comments:
1] I like strbracket:
'()' strbracket 'abc'
(abc)
2] If by 'tree' you mean J boxed lists of boxed lists to arbitrary
depths, here's a simple flattener:
wrote:
> strbracket =: {.@:[ , ] , {:@[
> joinB =: (1 : ' ((- # m)&}.@;@(,&m&.>"1))')(@: (] : ;))
> nums2name =:
