Apologies for top-posting... I have used Xerial successfully for a number of small-medium sized projects
http://www.xerial.org/trac/Xerial/wiki/SQLiteJDBC Contains precompiled libraries for win, linux and mac that are loaded automatically by JNI. It may still contain a pure java nestedvm translation for platforms that dont have native support. Can't really comment on its merits relative to other jdbc libraries, but it did everything I wanted. Add the jar to your software and off you go. Mike -----Original Message----- From: sqlite-users-boun...@sqlite.org [mailto:sqlite-users-boun...@sqlite.org] On Behalf Of Paul Linehan Sent: 27 April 2011 13:24 To: General Discussion of SQLite Database Subject: Re: [sqlite] Best JDBC driver for SQLite? Hi, Thanks for your reply. >> Gigs - what I'm interested in is sharding and using SQLite instances >> as shards in a larger ensemble. > Well now you've told us what you want out of it, we can comment usefully. Oh, goody! > You can handle your shards yourself, and you can implement sharding by > using VIEWs and TRIGGERs, but SQLite does not have sharding implemented /per se/. I know this - what I want to do is implement a sharding scheme myself with a queueing system and run it as a service on the different machines. These services will then communicate the result of any query back to the master server -> web server -> client. SQLite will be data storage only. I'll have to deal with the fact that writes in SQLite are blocking. The system I'm prototyping on is Firebird embedded, which is not client server either. Ideally, I want to use a serverless database system, but I will also be looking at Oracle XE and PosgreSQL. > If your design, from > the ground up, is based around multi-hosted data and/or client-server > architecture, then SQLite is not the system for you. I kind of want to do this for myself - check out my email address - it's a uni and this is a learning project as much as anything else. > You will be happier with MySQL or one of the other client/server > systems. I know I can do it this way - I could have a scheme where a truckload of machines run Firebird server (or MS SQL Server or Oracle or yadda yadda, but I'm trying to do it myself and take a minimalist approach. Perhaps the "shards" could be minimalist virtual machines running on the same box - if one tried to run 10 db servers at once on the same machine, the load would mount. What I have in mind is loosely based on Michael Stonebraker's VoltDB, except with a commodity off-the-shelf datastore. > However, there's no problem with the amount of data you specified, and > SQLite works on all your platforms (more than any other SQL DBMS I'm > aware of) without problems, since all the SQLite team supplies is the > source code for the function library. I'm doing this for my Java as much as anything else, so the platform must support that too, which is fine for Linux, *BSD and Windows. > If you need more details, ask more specific questions. Thanks again for your input. Rgs, Paul... > Simon. -- lineh...@tcd.ie Mob: 00 353 86 864 5772 _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
