delta/python-packages/rdflib.git, branch initbindings-fix github.com: rdflib/rdflib.git Fix initBindings handling. Fixes #294 2015-11-28T08:08:06+00:00 Gunnar Aastrand Grimnes gromgull@gmail.com 2015-11-28T08:08:06+00:00 6200934ace658ca05de693956b37623ce3a545eb This changes initBinding handling from adding values in the context object, to actually changing the query algebra and inserting a values clause in the right place. This also adds a bunch of tests. 
This changes initBinding handling from adding values in the context
object, to actually changing the query algebra and inserting a values
clause in the right place.

This also adds a bunch of tests.
Merge pull request #547 from gromgull/master 2015-11-22T13:57:35+00:00 Jörn Hees joernhees@users.noreply.github.com 2015-11-22T13:57:35+00:00 fa4172cc51a39b698ab7cb81240243ed6d62db1c Small but nice SPARQL Optimisation fix 
Small but nice SPARQL Optimisation fix
 cleaned up trailing whitespace 2015-11-22T13:13:38+00:00 Gunnar Aastrand Grimnes gromgull@gmail.com 2015-11-22T13:13:38+00:00 227b41fb1f1de15f1f4c126aae008444e0e6b0a9 






Small but nice SPARQL Optimisation fix
2015-11-22T08:40:45+00:00

Gunnar Aastrand Grimnes
gromgull@gmail.com

2015-11-20T19:21:11+00:00

cf9ccd9b894789cfbb90f341121f14bdb939c45e

This reorders the triples in a BGP while evaluating, so that the "most
bound" triples are evaluated first. The algebra transformation also does
this, but "statically", i.e. without looking at the data.

At runtime, we may have bindings from other places
(BIND/VALUES/initBindings) which can massively improve query time.

In a BerkelyDB store with ~250k triples, I had a query with two triples
patterns that retrieved 38 out of 25000 possible matching triples with a
particular property. Adding this fix brought query time from 8s to
800ms.





This reorders the triples in a BGP while evaluating, so that the "most
bound" triples are evaluated first. The algebra transformation also does
this, but "statically", i.e. without looking at the data.

At runtime, we may have bindings from other places
(BIND/VALUES/initBindings) which can massively improve query time.

In a BerkelyDB store with ~250k triples, I had a query with two triples
patterns that retrieved 38 out of 25000 possible matching triples with a
particular property. Adding this fix brought query time from 8s to
800ms.







Merge pull request #534 from lrowe/patch-2
2015-10-29T15:01:05+00:00

Jörn Hees
joernhees@users.noreply.github.com

2015-10-29T15:01:05+00:00

ea043ead8fd51acca7d2d3f4562aab7b05b2baa1

Avoid verbose build logging




Avoid verbose build logging






Avoid verbose build logging
2015-10-29T01:22:31+00:00

Laurence Rowe
l@lrowe.co.uk

2015-10-29T01:22:31+00:00

8cca20e096c579647520bd7636f88beacabda829












fixed a handful of stupid bugs in the csv2rdf tool
2015-10-27T11:13:30+00:00

Gunnar Aastrand Grimnes
gromgull@gmail.com

2015-10-27T11:11:36+00:00

ad1ec4e81381175da2c834f9f40e7d0411dbc2c0












Merge pull request #533 from joernhees/sparql_date
2015-10-24T12:26:00+00:00

Jörn Hees
joernhees@users.noreply.github.com

2015-10-24T12:26:00+00:00

18cce41ed5fdc7e4dc0189fc1107a8afc0e5c258

SPARQL can now compare xsd:date type as well, fixes #532




SPARQL can now compare xsd:date type as well, fixes #532






SPARQL can now compare xsd:date type as well, fixes #532
2015-10-24T12:09:19+00:00

Joern Hees
dev@joernhees.de

2015-10-24T12:09:19+00:00

a4002e3c502126cc82362e875fe61ce717912e6a












Merge pull request #529 from gromgull/master
2015-10-11T11:24:46+00:00

Jörn Hees
joernhees@users.noreply.github.com

2015-10-11T11:24:46+00:00

e1d0002fe699149865d8fa33937fa93806f4652a

let Resource objects be slices with other resources directly




let Resource objects be slices with other resources directly