| resizing an array of structures (uugh) [message #25234] |
Mon, 04 June 2001 04:57  |
Randall Skelton
Messages: 169
Registered: October 2000
|
Senior Member |
|
|
Hi all,
I have an ascii file containing a few thousand lines with each individual
dataset comprising 10 lines of floats, ints, strings, etc. It seems
logical to read this in as an array of structures as each dataset contains
the same information, just different numbers.
My problem is that I don't know what the dimension of the array should be
before I start. Initially I just defined a large array and counted the
number of datasets for subsequent processing. However, as time progresses
and this code gets more use, I have to say that I really hate all the
excess array elements... I figured there would be an easy way to resize
the array of structures, but the best I can come up with is a double for
loop that is rather slow.
; loop over the number of array elements
for i, n_elements(array) do begin
; loop over the number of tags
for j, n_tags(structure) do begin
resized_array[i].(j) = array[i].(j)
endfor
endfor
Is there a *faster* or more elegant way to do this? Does IDL have a
*fast* resize command that can handle any type of array to simply adjust
the number of elements in the array without rebinning, or otherwise
changing the numbers?
Cheers,
Randall
|
|
|
|
| Re: resizing an array of structures (uugh) [message #25337 is a reply to message #25234] |
Wed, 06 June 2001 10:34  |
Pavel A. Romashkin
Messages: 531
Registered: November 2000
|
Senior Member |
|
|
David Fanning wrote:
>
> Pavel A. Romashkin (pavel.romashkin@noaa.gov) writes:
>>
>> The experts answered already.
>>
>
> Pavel!! Where are you posting from? We had given you up
> for dead. It is nice to hear your voice again, even if there
> weren't any jokes. :-)
Since I haven't been using IDL much this last month, you can expect my
next code snippet to be the funniest joke of the month :-)
Cheers,
Pavel
P.S. I am back home, and that fresh, cold Sawtooth is soo-o easily
available ... where are you?
|
|
|
|
|
|
| Re: resizing an array of structures (uugh) [message #25341 is a reply to message #25234] |
Wed, 06 June 2001 10:01 |
Pavel A. Romashkin
Messages: 531
Registered: November 2000
|
Senior Member |
|
|
Hi Randall,
The experts answered already. I thought I might provide another way that
I found much faster than anything else for files up to 100,000 rows.
Read the file in a loop but use a large buffer, I use 10,000 rows. If
the size of the buffer exceeds the length of the file, use
Transfer_Count toobtain the right size for the buffer. Then, read the
leftover data. Below is an excerpt from my own routine. It is not
structure oriented but is easy to alter to fit your case:
; Read data from the file. Try to read by 10000 rows.
Buffer = fltarr(N_PrimVars+1) ; Because column 0 is always time.
readf, SourceFile, Buffer
; Create ResultArray with only 1 record, to initialize.
ResultArray = Transpose(Buffer)
; Create BIG buffer, in case file is long.
Buffer = fltarr(N_PrimVars+1, 10000)
; If Buffer was longer than file, reading will fail. In such case,
reduce buffer.
; Just to skip buffer reduction the first time,
FileTooShort = 0b
on_ioError, reduce_buffer
reduce_buffer : if FileTooShort eq 1b then begin
unread_rows = (fstat(SourceFile)).Transfer_Count / (n_primVars+1)
Buffer=fltarr(N_PrimVars+1, unread_rows)
point_lun, SourceFile, LastPos
endif
; In case of ioError, this will be set to 1b
FileTooShort = 1b
while not EOF(SourceFile) do begin
; Memorize start in case reading fails.
point_lun, (-SourceFile), LastPos
readf, SourceFile, Buffer
ResultArray = [ResultArray, Transpose(Buffer)]
endwhile
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
