That would be easy to do with a custom injector but personally I wouldn't want to make it a feature of the default injector. Where do we stop then? Accommodate all possible flavours of NoSQL, SQL etc... No thanks ;-) I am sure code like this could live happily on GitHub and use the Nutch maven artefacts. Such a custom injector could piggy-back the existing one, pull the seeds from MySQL into a temp crawlDB in a non-mapreduce fashion then reuse the second part of the injector which merges an existing crawldb with the freshly made one. Should be pretty straight forward to write I think HTH
Julien On 22 January 2013 17:12, Lewis John Mcgibbney <[email protected]>wrote: > This has certainly been explained in the past, however I can't find the > archived thread. > In short currently it is not possible. > I think it would be a nice feature for the injector though. > > On Tuesday, January 22, 2013, 刘兆贵 <[email protected]> wrote: > > Dear, > > I have a question, could you kind help me? Is there a way that I can > inject urls from a MySQL database rather than a text file. Thanks a lot! > > -- > *Lewis* > -- * *Open Source Solutions for Text Engineering http://digitalpebble.blogspot.com/ http://www.digitalpebble.com http://twitter.com/digitalpebble
