From: <tho...@us...> - 2014-01-10 17:26:03
|
Revision: 7760 http://bigdata.svn.sourceforge.net/bigdata/?rev=7760&view=rev Author: thompsonbry Date: 2014-01-10 17:25:56 +0000 (Fri, 10 Jan 2014) Log Message: ----------- javadoc edit. Modified Paths: -------------- branches/BIGDATA_RELEASE_1_3_0/bigdata-rdf/src/java/com/bigdata/rdf/sparql/ast/eval/AST2BOpJoins.java Modified: branches/BIGDATA_RELEASE_1_3_0/bigdata-rdf/src/java/com/bigdata/rdf/sparql/ast/eval/AST2BOpJoins.java =================================================================== --- branches/BIGDATA_RELEASE_1_3_0/bigdata-rdf/src/java/com/bigdata/rdf/sparql/ast/eval/AST2BOpJoins.java 2014-01-10 17:20:09 UTC (rev 7759) +++ branches/BIGDATA_RELEASE_1_3_0/bigdata-rdf/src/java/com/bigdata/rdf/sparql/ast/eval/AST2BOpJoins.java 2014-01-10 17:25:56 UTC (rev 7760) @@ -580,7 +580,7 @@ /* * TODO This optimization COULD be decided statically if we marked the * predicate with the index would would be used when it was evaluated. - * That is known in advance EXCEPT except when some joins are optional, + * That is known in advance EXCEPT when some joins are optional, * in which case the actual index can not be known until runtime. The * code which attaches the "as-bound" index to the predicate MUST also * consider the exogenous variables (if any). This might be done in the @@ -678,7 +678,7 @@ * cases where the PARALLEL SUBQUERY plan is faster than the * SCAN+FILTER. The approach coded here does not make the correct * decisions for reasons which seem to have more to do with the data - * density / sparsity for the APS which would be used for + * density / sparsity for the APs which would be used for * SCAN+FILTER versus PARALLEL SUBQUERY. Therefore the PARALLEL * SUBQUERY path for default graph access paths is currently * disabled. This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site. |