[kepler-dev] Some major fixes just checked in for Data Sources

Jing Tao tao at nceas.ucsb.edu
Thu Dec 16 11:35:34 PST 2004

Hi, Rod:

Thanks for the improvement. I tested some features so far so good.

In data source classes (both eml and digir) still has a vector - 
_dataVector to store the data. It will cause problem when data file is 
huge. I am think to remove the vector and replaced by the inputstream. How 
do you think?


On Thu, 16 Dec 2004, Rod Spears wrote:

> Date: Thu, 16 Dec 2004 12:48:36 -0600
> From: Rod Spears <rods at ku.edu>
> To: kepler list <kepler-dev at ecoinformatics.org>
> Subject: [kepler-dev] Some major fixes just checked in for Data Sources
> 1) I changed the name of an XML element that had been wrong for a while 
> (returnfield -> returnField)
> this means you should delete your cache.
> 2) The caching mechanism and the Data Sources no longer use strings for 
> passing cached data (we knew this would be a bad thing at some point), 
> so now the parsers parse directly from the local data file in the cache. 
> In the future we can switch it to use streams, but since a lot of the 
> parsing code was laready using helper function function which could 
> easily take a file name I did that instead. It will only byte us if we 
> switch away from the current cache implementation where we use 
> individual file names, wwich at the moment is likely because indv. file 
> names work well HSQL. NOte because of the largeness of the DiGIR 
> datasets they were not being obtinaed from the cache correctly, so now 
> we are getting the cached resultsets instead of refetching them
> So update everything, rebuild and delete your cache.
> Rod
> _______________________________________________
> kepler-dev mailing list
> kepler-dev at ecoinformatics.org
> http://www.ecoinformatics.org/mailman/listinfo/kepler-dev

Jing Tao
National Center for Ecological
Analysis and Synthesis (NCEAS)
735 State St. Suite 204
Santa Barbara, CA 93101

More information about the Kepler-dev mailing list