| Re: Another strsplit question [message #37747] |
Wed, 21 January 2004 08:31  |
kneller
Messages: 4
Registered: January 2004
|
Junior Member |
|
|
MKatz843@onebox.com (M. Katz) wrote in message news:<4a097d6a.0401201249.3299ca9b@posting.google.com>...
> kneller@nist.gov (Larry Kneller) wrote in message news:<96e91d8d.0401200800.34fa60fa@posting.google.com>...
>> Most functions in IDL will operate on arrays.
>> Is there a way to make strsplit allow this?
>>
>> e.g.
>> s = ['I see the computer.','I am typing a message.',$
>> 'This is the idl-pvwave newsgroup.']
>>
>> segs = strsplit(s,/extract,/some_option)
>>
>> print,segs[0,0],segs[1,0],segs[2,0],segs[2,4]
>>
>> RESULT WOULD BE:
>>
>> I I This newsgroup.
>
> The difficulty here is that since every input string is different,
> every row of the output array will want to have a different number of
> values. So it's ambiguous how you'd write it. Would you look for the
> row with the most elements and make all rows be that long? That's one
> way. Or would you have the output be a pointer array where each
> element points to an array of strings. That's probably more memory
> efficient in the long run, but some people might not like dealing with
> pointers, and you'd run the risk of asking for array elements that
> don't exist.
>
> Looks like you need to write your own function here to fit your needs.
> That's pretty straightforward to do in a case like this.
>
> M. Katz
Thanks for the suggestions! I think the array of pointers makes more
sense for my personal use. That would easily allow a ragged array
of 'strings'.
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
