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

[netCDF #FQQ-870725]: NetCDF use



The first thing you should do is
write a CDL file describing your
data. You can use the ncgen utility
to check the correctness of that
CDL description.

Further, you should use netcdf-4
and not the more limiting netcdf-3.

Now, here is my first try at answering
your question:

1. How do I store the patient’s name? Is that possible?

You can store a name as a variable of type String.

2. I’ve mapped every patient to a NetCDF dimension. To retrieve a
patient’s data I need to map the patient to its dimension. Can this
information be kept on NetCDF or do I need to store this mapping on a
database outside NetCDF? For instance, John Doe is mapped to dimension
25 (my 25th patient) on the data.

One simple way to do this is to define a variable like this:
    dimensions:
        numberofpatients=2000;
    variables:
        String names(numberofpatients);
Then you can search the names variable for a matching patient name.
The index of names at which you find the matching name then
becomes the index of the patient.

3. I use time as an unlimited dimension, with time 0, as the first day
of the program. When a new patient enters the program, we do not have
any data for her yet. Let’s say that she enters the program after 100
days, if we use a time index of 100 NetCDF will create empty slots for time
0 to 99 for this patient and if I understand correctly, this will waist
space. Is that the right way to do this?

netcdf-3 will waste this space. If you use netcdf-4
(as I recommend), then the space is managed more efficiently
internally and no space is wasted.

4. On the same line of question 3, when I stop collecting “normal data”
to collect “treatment data”, there will be a gap on the data. For
instance, let’s say that I stop collecting normal data on day 100 and
resume it on day 150, will there be a gap of 50 days with empty records
from day 100 to day 149?

Same answer as #3. But I would not worry about the wasted space
unless you really think your data will have a really large
number of records.  The space is wasted on the disk (with netcdf-3)
but does not affect memory usage.

=Dennis Heimbigner
 Unidata


=Dennis Heimbigner
  Unidata


Ticket Details
===================
Ticket ID: FQQ-870725
Department: Support netCDF
Priority: High
Status: Closed