[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: proper netcdf creation practice

Hi Tim,

> Date: Wed, 19 Nov 2003 15:01:19 -0700 (MST)
> From: Tim Hoar <address@hidden>
> To: <address@hidden>
> Subject: proper netcdf creation practice

In the above email, you wrote:

> Can you point me to the example of the proper way to dynamically append
> to the unlimited dimension?
> 
> My writing gets much slower as the length of the unlimited dimension
> (in this case, time) gets larger.
> 
> I poked around on the unidata netcdf www-site with no success.

Here's the message on the netcdfgroup mailing list that showed the
time taken to append data to a large file depends on the amount of
data, not the size of the file:

  http://www.unidata.ucar.edu/cgi-bin/msgout?/glimpse/netcdfgroup-list/1455

and the C program referenced there is still available from

  http://www.unidata.ucar.edu/packages/netcdf/jg1.c

This was in a response to a complaint that it was taking longer and
longer to append data to a file as it grew larger, but that turned out
to be a bug in the Python interface to netCDF, not in the netCDF
library:

  http://www.unidata.ucar.edu/cgi-bin/msgout?/glimpse/netcdfgroup-list/1458
  http://www.unidata.ucar.edu/cgi-bin/msgout?/glimpse/netcdfgroup-list/1453

Here was another response that independently showed no slowdown:

  http://www.unidata.ucar.edu/cgi-bin/msgout?/glimpse/netcdfgroup-list/1454

The whole thread of discussion, from most recent to oldest, is
available here:

  
http://www.unidata.ucar.edu/cgi-bin/aglimpse/82?query=degrades&case=on&whole=on&errors=0&maxfiles=20&maxlines=10

I hope this helps ...

--Russ