| indexing arrays with arrays [message #55222] |
Tue, 07 August 2007 06:39  |
Conor
Messages: 138
Registered: February 2007
|
Senior Member |
|
|
It seems to me that a highly useful syntax for IDL would be something
like this:
res = randomu(seed,10)
start_ind = indgen(5)
end_ind = start_ind + 1
extract = res[start_ind:end_ind]
In this case, I would envision extract being a 10 element array.
Namely, it would be the equivelent of:
extract = [res[start_ind[0]:end_ind[0]], res[start_ind[1]:end_ind[1]],
res[start_ind[2]:end_ind[2]], etc...]
Does anyone else think something like this would be useful? Anyone
know how I can bug RSI and request something like this? I couldn't
find a "feature request" option, or anything similar, on their site.
Anyway, that point aside, can anyone think of a robust way to do this
right now? I've been trying to come up with a speedy, robust, for-
loop-less way to do this, but haven't come up with anything. Any
thoughts would be appreciated.
|
|
|
|
| Re: indexing arrays with arrays [message #55252 is a reply to message #55222] |
Thu, 09 August 2007 09:45  |
JD Smith
Messages: 850
Registered: December 1999
|
Senior Member |
|
|
On Tue, 07 Aug 2007 06:39:54 -0700, Conor wrote:
> It seems to me that a highly useful syntax for IDL would be something
> like this:
>
> res = randomu(seed,10)
> start_ind = indgen(5)
> end_ind = start_ind + 1
>
> extract = res[start_ind:end_ind]
>
> In this case, I would envision extract being a 10 element array.
> Namely, it would be the equivelent of:
>
> extract = [res[start_ind[0]:end_ind[0]], res[start_ind[1]:end_ind[1]],
> res[start_ind[2]:end_ind[2]], etc...]
>
> Does anyone else think something like this would be useful? Anyone
> know how I can bug RSI and request something like this? I couldn't
> find a "feature request" option, or anything similar, on their site.
> Anyway, that point aside, can anyone think of a robust way to do this
> right now? I've been trying to come up with a speedy, robust, for-
> loop-less way to do this, but haven't come up with anything. Any
> thoughts would be appreciated.
Well, if end_ind is always a constant offset from start_ind, it's easy
just to contruct the index vector yourself:
t=[end_ind[0]-start_ind[0]+1,n_elements(start_ind)]
extract=res[reform(rebin(1#start_ind,t)+rebin(indgen(t[0]),t ),t[0]*t[1])]
If the start-end difference can be any variable amount, this is a
problem for HISTOGRAM related to chunk indexing. See the HISTOGRAM
tutorial and:
http://www.dfanning.com/idl_way/chunkindex.html
JD
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
