Re: choosing a lightweight database

[John Nemeth <jnemeth%cue.bc.ca@localhost>]

On Jan 16, 11:03am, Greg Troxel wrote:
} reed%reedmedia.net@localhost writes:
} 
} > Around 200 attributes, but not all for each.
} >
} > Also I have key/values like
} >  FOO-notes="for supplementing text for FOO"
} > And description and direction fields with sentences.
} >
} > The values may be free form text, numbers, boolean yes/no (some values 
} > are links to other files).
} >
} > I generate static webpages easily from this (and can dynamically 
} > generate webpages easily).
} >
} > Storing this in Berkeley DB would be easy.
} >
} > But I want an easy way to search everything like:
} >
} > country=ca
} > province=alberta
} > elevation>=1524 meters
} > cost<=5
} >
} > or keyword searches against description fields.
} 
} If you don't want a db process, I would try sqlite and as long as it
} holds up stick with it.
} 
} If it turns out your data size or query/update rate is too much, I would
} use postgres.  I know you said you don't want a process, but unlike
} mysql postgres is really easy to set up.

     It would be nice to get rid of some of the FUD around here.
MySQL is quite simple to setup, especially if you're installing it
from pkgsrc.  mysql-cluster is complex to setup, but that is for
master-master replication with redundancy.  The regular mysql-server
is just pkgin mysql-server, set the "root" password and you're off
to the races.

} And, it sounds like you should be contributing to openstreetmap :-)
}-- End of excerpt from Greg Troxel