Oh... so many questions. But first, many, many thanks Craig for
sending this module. I will try it out.

On Fri, Aug 27, 2010 at 1:03 AM, Craig DeForest <[email protected]> wrote:
> Whups - forgot to include it.  Here.
>
...
> On Aug 26, 2010, at 11:46 PM, Craig DeForest wrote:
>
>> Why not tell him to write out all the data in a .sav file, and read it in
>> using PDL::IO::IDL?

1. I don't see this module on the PDL web site? Is it not for public
consumption?

2. He is working on creating the .sav files, but hasn't done so yet.
Will test it out soon as he is ready. Part of the problem is, he is
experiencing some problems on the IDL side. If he writes out 503 x 539
x 100 array, he gets a 108446800 bytes in his binary file.  I am able
to read those files in PDL properly (although there are still data gap
problems). However, if he writes out a file with 271117 x 100 array
(that is, collapses 503 x 539 into a single dim), he gets 108445696
bytes in his binary file. That is, of course, 1104 bytes less than how
much the file should be (27111700 * 4 bytes for floats). Any idea why
that might be happening? I know this is a long shot... why would any
one of you know what is wrong with my colleague's program and output,
but hey, you all are brainiacs and mind-readers... so...

3. I am trying to read up on PDL datatypes, but methinks, those docs
need to be clarified and organized better. For example, I can't seem
to find the definitions and sizes of each of the PDL datatypes. The
first hit I get just tells me what they are... byte, short, long,
float, double, etc., but not how much space they take. So, what is the
difference between a float and a double, and a long? I can wikipedia
this stuff, but it really belongs write there in PDL docs, at least,
imho.

The piddle I created yesterday from the text file ended up being a
Double, and was, of course, double in size from that of a float!

Nevertheless, any idea why my avgs are turning out to be nans?



>>
>> On Aug 26, 2010, at 11:41 PM, P Kishor wrote:
>>
>>> I am running into a strange problem converting from data coming from
>>> IDL. A colleague is generating the IDL output as 100 values each for
>>> 271117 cells. This is what I would think of as a 271117 x 100 array of
>>> arrays. That is, an array with 271K elements, each element being a 100
>>> element array.
>>>
>>> Except, when I read into PDL via FlexRaw, using a header of "Float
>>> 271117 100", I get the error "PDL: Couldn't read enough data from
>>> <file name>..." Strangely, when I change the header to 271116, 99, it
>>> works, but the output is all wrong... lots of gaps. I can't figure out
>>> why.
>>>
>>> Ok. To get to the bottom of this mystery, I asked him to write out all
>>> the values as a text file so I could see the darned data. Well, he
>>> didn't know how to format text output from IDL so he wrote out one
>>> value per line and gave me a text file with 27111700 (27 million
>>> lines... please control your snickering and laughter).
>>>
>>> I coaxed that file into a PDL which looks like so
>>>
>>>  PDL: Double D [100,271117]
>>>
>>> The header generated by FlexRaw looks like so
>>>
>>>  Double
>>>  2
>>>  100
>>>  271117
>>>
>>> So, I read this piddle back into PDL. This is what it looks like
>>>
>>>  print $pdl->info . "\n";
>>>>
>>>> PDL: Float D [100,271117]
>>>
>>> Now I want to average the numbers along the 0th dimension (those are
>>> 100 years), that is, I want yearly averages for all the cells. So I do
>>> the following
>>>
>>>  my $xpdl = $pdl->xchg(0, 1);
>>>  my $yavgs = $xpdl->average;
>>>  print $yavgs->info . "\n";
>>>>
>>>> PDL: Float D [100]
>>>
>>> Now I print the entire avg piddle, and I get garbage.
>>>
>>> [nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan
>>> nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan
>>> nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan
>>> nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan
>>> nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan
>>> nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan]
>>>
>>> Any idea what I could be doing wrong?
>>>



-- 
Puneet Kishor http://www.punkish.org
Carbon Model http://carbonmodel.org
Charter Member, Open Source Geospatial Foundation http://www.osgeo.org
Science Commons Fellow, http://sciencecommons.org/about/whoweare/kishor
Nelson Institute, UW-Madison http://www.nelson.wisc.edu
-----------------------------------------------------------------------
Assertions are politics; backing up assertions with evidence is science
=======================================================================

_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl

Reply via email to