Hello!
Referring to:
"If the error is suppressed (with "OnError Resume Next") then I can read
some simple data from the *.cdf file. But if string variables or array
variables are involved, the program fails or even crashes."
for this case, I want to right away say that you need to keep an eye on how you
pass your arrays. You have to pass them by referencing the first element of
the array. For example, I want to load a longitude variable into an array:
'ncFile is the handle to a NetCDF file from nc_open
'lon_id is the longitude variable id from nc_inq_varid
'lon_length is the size of the longitude dimension
Redim lon_array!(1 To lon_length&)
NcErr& = nc_get_var_float(ncFile&, lon_id&, lon_array!(1))
Same thing for a 2D array:
Redim lon_array2!(1 To cellx_length&, 1 To celly_length&)
NcErr& = nc_get_var_float(ncFile&, lon_id&, lon_array2!(1, 1))
For strings, you must make sure the string variable you are passing in has
space already allocated for it by using Space or String. VB takes care of
passing it your function properly as long as your string parameter is
declared to be ByVal in the nc_get_att_text declaration. You can determine
the size of the string by looking at the dimensions of the variable. For
attributes you would use nc_inq_attlen or nc_inq_att. Once you have the
string length you must create a string of the right size to contain the
string data. For example, I want to get the units for the longitude
variable above:
NcErr& = nc_inq_attlen(ncFile&, lon_id&, "units", units_len&)
units$ = String(units_len& + 1, Chr$(0))
NcErr& = nc_get_att_text(ncFile&, lon_id&, "units", units$)
'Can precede units$ with ByVal
I got the string length for the attribute, created a string of Null characters
of the appropriate length (I like to include an extra character just in case),
and then I retrieved the string from the attribute.
In summary, make sure you have memory room in your arrays and strings since
the NetCDF API doesn't allocate those for you. Make sure you specify the first
indices of the array when you are retrieving a variable into an array. And,
finally, make sure your declarations are correct. Where the C API says byte
use integer, short use integer, int use long, float use single, and double use
double. Be sure your declaration uses ByVal when passing strings.
Good Luck!
Matthew Hanna
ncdigest wrote:
ncdigest Thursday, December 9 2004 Volume 01 : Number 828
Today's Topics:
netcdf.dll and Visual Basic 6
----------------------------------------------------------------------
Date: Thu, 9 Dec 2004 19:27:08 +0100 (MET)
From: =?ISO-8859-1?Q?=22Hans-J=FCrgen_Rieger=22?= <H.Rieger@xxxxxx>
Subject: netcdf.dll and Visual Basic 6
Hi,
I´ve got some problems using the netcdf.dll (prebuilt dll from the
netcdf-3.5.1 download) with Visual Basic 6.
In a new module I wrote the declaration:
Declare Function nc_open Lib "netcdf.dll" (ByVal path As String, ByVal
cmode As Long, ByRef ncidp As Long) As Long
Then I used it in a procedure:
Sub main
Dim path As String, ncidp As Long, res As Long
path = "c:\test.cdf"
res = nc_open(path, 0, ncidp)
End Sub
Running the program, I get the error message "Bad dll calling convention"
If the error is suppressed (with "OnError Resume Next") then I can read
some simple data from the *.cdf file. But if string variables or array
variables are involved, the program fails or even crashes.
I´ve got the impression from other netcdfgroup members that it is
possible to use the
netcdf.dll in the way outlined above.
Does anybody know what I am doing wrong? What calling convention does
the netcdf.dll uses?
With best regards
H.J. Rieger
- --
GMX ProMail mit bestem Virenschutz http://www.gmx.net/de/go/mail
+++ Empfehlung der Redaktion +++ Internet Professionell 10/04 +++
------------------------------
End of ncdigest V1 #828
***********************
From owner-netcdfgroup@xxxxxxxxxxxxxxxx 10 2004 Dec -0700 09:23:17
Message-ID: <wrxr7lydswa.fsf@xxxxxxxxxxxxxxxxxxxxxxx>
Date: 10 Dec 2004 09:23:17 -0700
From: Ed Hartnett <ed@xxxxxxxxxxxxxxxx>
In-Reply-To: <417432E2.1010500@xxxxxxxxxx>
To: Mark Hadfield <m.hadfield@xxxxxxxxxx>, Harald.Anlauf@xxxxxx,
Subject: Re: netCDF 3.6.0-beta6 (almost) success on cygwin+g95
Received: (from majordo@localhost)
by unidata.ucar.edu (UCAR/Unidata) id iBAGNT0g015598
for netcdfgroup-out; Fri, 10 Dec 2004 09:23:29 -0700 (MST)
Organization: UCAR/Unidata
Keywords: 200412101623.iBAGNIlI015586
netcdfgroup@xxxxxxxxxxxxxxxx
CC: russ@xxxxxxxxxxxxxxxx
References: <52704F189EF3D811910900055DE6449E4A604D@ofexc0>
<417432E2.1010500@xxxxxxxxxx>
Lines: 42
User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Sender: owner-netcdfgroup@xxxxxxxxxxxxxxxx
Precedence: bulk
Reply-To: Ed Hartnett <ed@xxxxxxxxxxxxxxxx>
Mark Hadfield <m.hadfield@xxxxxxxxxx> writes:
Anlauf Harald wrote:
>first of all: Congratulations! Building and testing this beta went
>very smooth on my cygwin system with a self-built g95 compiler from
>www.g95.org .
>
>The only problem I encountered was due to a bug in the install script
>that did not expect that the ncdump and ncgen executables have an
>.exe ending on cygwin:
>
>...
>A possible solution for Windows/Cygwin systems: define a variable
>EXEEXT in the top-level Makefile which is set to EXEEXT by configure
>(and left empty for real Unix systems), and fix the Makefiles in the
>ncgen and ncdump directories accordingly.
>
A better solution: use the "install" command. On Cygwin this command
copes with the .exe extension automatically.
Is there any reason not to modify the standard netCDF makefiles to use
"install" in place of "cp"? I am willing to submit a patch.
--
Mark Hadfield "Ka puwaha te tai nei, Hoea tatou"
m.hadfield@xxxxxxxxxx
National Institute for Water and Atmospheric Research (NIWA)
Mark and Anlauf,
Thanks for pointing out the problem installing under cygwin. I have
switched to using install instead of cp, and this change will be
included in the 3.6.0 release.
If you have any other suggestions to make, I would be very happy to
learn of them.
Thanks again,
Ed Hartnett
Unidata
