RE: [boost] bdb and mpi selectors for the graph library
Hi. I don't understand what do you propose.
The second is aimed at situations when you can't distribute your graph (e.g. you're partitionning it before distributing it), or speed isn't an issue. It uses a berkeley DB database to store the graph.
Do you propose interface that suitable for boost::graph library while implementation of graph is BDB ? If so I'd like to see some code or\and documentation. Roman
What I propose, in the case of bdb, is just an edge and a vertex container, which allows you to do something like : typedef adjacency_list<bdbS, bdbS, ...> Graph; and store you whole graph in a berkeley DB database, so that you can play with fairly big graphs... Hope it answers your question... Pierre Le 3 mai 04, à 16:06, Roman Yakovenko a écrit :
Hi. I don't understand what do you propose.
The second is aimed at situations when you can't distribute your graph (e.g. you're partitionning it before distributing it), or speed isn't an issue. It uses a berkeley DB database to store the graph.
Do you propose interface that suitable for boost::graph library while implementation of graph is BDB ? If so I'd like to see some code or\and documentation.
Roman
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
-- Written slowly for people who can't read fast.
Why you choose Berkley DB in particular? Mohammed ----- Original Message ----- From: "Pierre DOUCY" <pdoucyml@free.fr> To: <boost@lists.boost.org> Sent: Tuesday, May 04, 2004 10:52 PM Subject: Re: [boost] bdb and mpi selectors for the graph library
What I propose, in the case of bdb, is just an edge and a vertex container, which allows you to do something like : typedef adjacency_list<bdbS, bdbS, ...> Graph; and store you whole graph in a berkeley DB database, so that you can play with fairly big graphs...
Hope it answers your question...
Pierre
Le 3 mai 04, à 16:06, Roman Yakovenko a écrit :
Hi. I don't understand what do you propose.
The second is aimed at situations when you can't distribute your graph (e.g. you're partitionning it before distributing it), or speed isn't an issue. It uses a berkeley DB database to store the graph.
Do you propose interface that suitable for boost::graph library while implementation of graph is BDB ? If so I'd like to see some code or\and documentation.
Roman
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
-- Written slowly for people who can't read fast.
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
participants (4)
-
Mohammed Iqbal -
Pierre DOUCY -
Roman Yakovenko