Bulk data load into OQGRAPH
Is there any batch mode mechanism for loading a lot of data into OQGRAPH? I have an application that has about 11 million edges and I want to be able to load it up quickly. I created an output file that contains the first 1 million edges to use for testing. I then tried using
mysql> LOAD DATA INFILE 'oqgraph_1M.out' INTO TABLE graph (origid, destid, weight);
It took exactly 70 minutes to load these 1 M rows. I did not try to load the entire 11 M rows yet, because that seems like it would take at least 11 times longer (i.e. at least 12.8 hours).
Is there some other mechanism that would be faster to do a bulk load? The problem is that since the graph engine does not have any persistence mechanism, then every time a database is stopped and restarted (e.g. for planned maintenance, etc), then the engine will have to be reloaded from the persistant table, which will take at least 13 hours.
I'm not sure what the bottleneck is for the loading, is it updating the index? If so, is there a way to defer that until all the data is loaded?