[CF-metadata] stations and trajectories

From: John Caron <caron>
Date: Mon, 20 Jun 2005 14:12:49 -0600

Hi Jonathan:

Jonathan Gregory wrote:

>Dear John
>
>I don't think we can arrange unique station names unless we define or point
>to a list of them somehow. Should we do this?
>
>
My solution would just be to tell the writer that the names must be
unique, not to specify what they are.

>
>
>>this has a good feel to it, you could elaborate on this like:
>>
>>float measure1( record);
>> :coordinates = "time lat lon elev station";
>>double time( record);
>>double lat( record);
>>double lon( record);
>>double elev( record);
>>int station( record);
>>
>>
>
>This is like storing all the timeseries in a single SQL table, when you would
>extract the timeseries for a particular station using
> select time,measure1 where station=something;
>Is that how you did it in SQL, Magi?
>
>I understand you would like to do it like this in order not to have many
>variables and I agree this is a possible method. But why is it preferable to
>having many variables? Is it because it is easier to search through a variable
>to find a particular value (of station, for instance) than it is to search
>through a number of variables to identify the one you want? If that's the
>reason there might be other approaches we could think of.
>
>
Our metar files that we produce from realtime data have 4000 stations in
them. If each station needed its own time coordinate and dimension, well
that would be an ugly mess!
Received on Mon Jun 20 2005 - 14:12:49 BST