Java woes...
My "unexpected error" Java problems haven't been properly resolved yet, but in the meantime, I've found a workaround by splitting up the database into several parts based on continents. So, to start with, I have one database of all places in Europe using Berkeley Java DB.
Doing lookups in this database is pretty quick - on a 2.6GHz Intel Celeron processor with 512Mb RAM, looking up 200+ names takes around 500 ms. Not too shabby!
Initially I decided against serializing objects as the values in the database table, instead using strings with embedded field delimiters so I could use regexps to split them up later. But which approach would lead to faster access to individual elements once lookup has been done and also which would lead to a smaller database? So this is one experiment that has to be done.
Doing lookups in this database is pretty quick - on a 2.6GHz Intel Celeron processor with 512Mb RAM, looking up 200+ names takes around 500 ms. Not too shabby!
Initially I decided against serializing objects as the values in the database table, instead using strings with embedded field delimiters so I could use regexps to split them up later. But which approach would lead to faster access to individual elements once lookup has been done and also which would lead to a smaller database? So this is one experiment that has to be done.
posted by Simon at 6:09 PM
0 Comments:
Post a Comment
<< Home