The ComplexPhraseQueryParser now has some features, including normalization of multiterms, that didn't exist when I created the SpanQueryParser. Given that SpanQueries are buggy, we should allow users to select their queryparser.
We should also allow users to select the classic queryparser so that they can transfer queries developed on Rhapsode to another Lucene-based engine without problem.
The queryparser information will need to be stored with each stored query (and concept?!).
If a user selects the classic queryparser/CPQP, make sure to pre-parse with the SpanQueryParser to identify features not available in the classic query parser. That is: warn people that SpanQueryParser functionality isn't available if they try to use a different parser. Currently, there is no warning, just a failed query.
The ComplexPhraseQueryParser now has some features, including normalization of multiterms, that didn't exist when I created the SpanQueryParser. Given that SpanQueries are buggy, we should allow users to select their queryparser.
We should also allow users to select the classic queryparser so that they can transfer queries developed on Rhapsode to another Lucene-based engine without problem.
The queryparser information will need to be stored with each stored query (and concept?!).
If a user selects the classic queryparser/CPQP, make sure to pre-parse with the SpanQueryParser to identify features not available in the classic query parser. That is: warn people that SpanQueryParser functionality isn't available if they try to use a different parser. Currently, there is no warning, just a failed query.