| Re: IDL: are dynamic arrays supported? [message #10878] |
Wed, 11 February 1998 00:00 |
davidf
Messages: 2866
Registered: September 1996
|
Senior Member |
|
|
Darran Edmundson (dEdmundson@Bigfoot.com) writes:
> I am a new user of IDL. I am trying to input data for a 2d surface
> created by a Fortran program. The data has the following form:
>
> xnum
> x1 x2 x3 x3 x5
> x6 x7 x8 ...
> y1
> z1 z2 z3 z4 z5
> z6 z7 z8 ...
> y2
> z1 z2 z3 ...
>
> Here, xnum is the number of x gridpoints, x1..xn form the x(xnum) vector,
> y1, y2, ... form the y(ynum) vector and the z1..zn are the height data
> at the current y value. The problem is that ynum isn't explicitly
> listed at the beginning of the file.
>
> The obvious solution is to read the file twice, first to glean ynum,
> and second to actually grab the data. However, I wonder is it possible
> (and desirable) to import the data in a single pass? My IDL manual
> suggests creating a 2D array bigger than I think I will need. This is
> inelegant and I wonder about some form of dynamic array.
Humm. I am notoriously bad at this kind of analysis, but I'm
sitting here with my leg up in the air popping pain pills
for my bum knee, so I figure this time I have a built-in
excuse: my brain is addled.
I don't believe this file can be read in the manner described
in the IDL manuals. They refer to a 2D array of values, when
what you really have here, it seems to me, are X and Y vectors
and the 2D array that corresponds to values at the
locations specified by the vectors, with the file arranged
in a slightly obscure way.
I presume that although the specific number of Y elements is
not written in the file, it is a specific known number. If
this is not the case, then I can't think of a way to avoid
reading the data in some kind of loop. Perhaps like this:
ReadF, lun, newY, ZVector
y = [Temporary(y), newY]
z = [Temporary(z), ZVector]
Where you then reform Z into the proper sized 2D array at the
end of the loop. This is the sort of "dynamic array" concept
that you mention.
But suppose the number of Y elements were known. Say it is 30.
Then I might read this dataset with code like this:
; Read the X vector
ReadF, lun, xnum
x = FltArr(xnum)
ReadF, lun, x
; Read the rest of the data in one gulp. Assume ynum=30.
data = FltArr(30 * xnum)
ReadF, lun, data
; Reform the data so that the y vector is the first column.
data = Reform(data, xnum+1, 30)
; Separate the y vector from the z data.
y = Reform(data[0,*])
z = Temporary(data[1:*,*])
That should do it. This would be considerably faster than a loop.
And at least in my addled state, it *seems* elegant. :-)
Cheers,
David
|
|
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
