Comments on: GovTrack RDF Data http://crschmidt.net/blog/58/govtrack-rdf-data/ Ramblings of a GIS Hacker Thu, 04 Dec 2008 02:57:03 +0000 http://wordpress.org/?v=2.6.3 By: Joshua Tauberer http://crschmidt.net/blog/58/govtrack-rdf-data/#comment-1811 Joshua Tauberer Tue, 26 Apr 2005 18:02:20 +0000 http://crschmidt.net/blog/archives/58/govtrack-rdf-data/#comment-1811 Neat! Don't know if you've seen <a href="http://www.govtrack.us/rdfbrowse.xpd">this</a>. It helps for (mentally) visualizing the data. (And, BTW, the data gets *much* more interesting once the N3 files are loaded, although it takes me a good few hours to get it all into MySQL.) Neat! Don’t know if you’ve seen this. It helps for (mentally) visualizing the data. (And, BTW, the data gets *much* more interesting once the N3 files are loaded, although it takes me a good few hours to get it all into MySQL.)

]]> By: Christopher Schmidt http://crschmidt.net/blog/58/govtrack-rdf-data/#comment-1804 Christopher Schmidt Tue, 26 Apr 2005 11:37:49 +0000 http://crschmidt.net/blog/archives/58/govtrack-rdf-data/#comment-1804 Hm... but the joins are against the node tables, no? It still has to look up the indexes in those Node tables. I suppose that's not much time unless, like me, you've got indexes that are larger than your keycache. Then you have to fall back to disks, which are impossibly slow. I suppose everything is indexed six ways from Sunday, so it's really a moot point for most things. I just have a slow computer and like dealing with impossibly large datasets :) So, starting in an existing Model is slow, a new Model is faster, and a new Database even faster. Hm… but the joins are against the node tables, no? It still has to look up the indexes in those Node tables. I suppose that’s not much time unless, like me, you’ve got indexes that are larger than your keycache. Then you have to fall back to disks, which are impossibly slow.

I suppose everything is indexed six ways from Sunday, so it’s really a moot point for most things. I just have a slow computer and like dealing with impossibly large datasets :)

So, starting in an existing Model is slow, a new Model is faster, and a new Database even faster.

]]> By: Morten Frederiksen http://crschmidt.net/blog/58/govtrack-rdf-data/#comment-1803 Morten Frederiksen Tue, 26 Apr 2005 10:16:51 +0000 http://crschmidt.net/blog/archives/58/govtrack-rdf-data/#comment-1803 Re #3: That's not quite right. The select issued to check for existing statements only involve the StatementsX table, not the node tables. Thus, loading into a new model will be faster. Re #3:

That’s not quite right. The select issued to check for existing statements only involve the StatementsX table, not the node tables. Thus, loading into a new model will be faster.

]]>