On 7/15/06, Rob Staveley (Tom) <[EMAIL PROTECTED]> wrote:
The problem one is testSortingReverseNullFirst. My reckoning is that you are forward-sorting (or ascending-sorting if that's the right thing to say) the non-nulls.
OK, examples are probably easier to work with. If we have a list of [A,null,Z] and we want that reverse sorted, then we want nulls first, the result should be [null,Z,A]. If we were to use nullStringLastComparatorSource (which orders nulls last in an *ascending* Lucene sort), the result would be [A,Z,null]. If we reverse this lucene sort, we get [null,Z,A], the desired result. Ahhh, but in the code, we aren't doing the last reverse! That does look wrong. return new SortField(fieldName,nullStringLastComparatorSource); // <-- !!! Great catch! I think we never noticed because all of our usecases called for nulls to sort last. I'll add a test to Solr for this case. -Yonik http://incubator.apache.org/solr Solr, the open-source Lucene search server --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
