Thanks, Andy. I'm running you scripts and the following errors prompt. Does it mean the error was fixed?
9 ==== data/xaa (Mon Aug 26 09:30:02 CST 2013) 10 ERROR [line: 4692554, col: 35] Unknown char: \(92) 11 stdin : 42.79 sec 4,692,540 triples 109,656.72 TPS 12 ==== data/xab (Mon Aug 26 09:30:46 CST 2013) 13 stdin : 86.39 sec 9,999,991 triples 115,751.35 TPS 14 ==== data/xac (Mon Aug 26 09:32:13 CST 2013) 15 ERROR [line: 2921254, col: 41] Unknown char: \(92) Thanks, Tao -----Original Message----- From: Andy Seaborne [mailto:[email protected]] Sent: Friday, August 23, 2013 12:14 AM To: [email protected] Subject: Re: Has anyone loaded freebase dump to TDB successfully? On 22/08/13 10:57, Tao (陶信东) wrote: > I used to try but failed (due to some format errors). Now I want to > try again? > > > > Do I have a chance to succeed? Yes - there's chance. But. First, you need to make sure the data is clean - that is, it parses. http://people.apache.org/~andy/Freebase20121223/Notes.txt has my notes from that published version. Theer is no point loading until the data parses cleanly. Use "riot --validate" Then you need a big (a lot of RAM) 64 bit machine and JVM and patience. It's a large load and TDB slows down at scale currently. There have been reports of exceptionally good loading, but it was on a large, powerful machine. More RAM = faster. The loading may well take more a very long time. Andy > > > > Thanks > > Tao > >
