[
https://issues.apache.org/jira/browse/LUCENE-2458?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12866696#action_12866696
]
Hoss Man commented on LUCENE-2458:
----------------------------------
bq. Instead the queryparser should only form phrasequeries when you use double
quotes, just like the documentation says.
i'll grant you that the documentation is wrong -- i view that as a bug in the
documentation, not the code.
Saying that a PhraseQuery object should only ever be constructed if the user
uses quotes is like saying that a BooleanQuery should only ever be constructed
if the user specifies boolean operators -- there is no rule that the syntax
must match the query structure, the same Query classes can serve multiple
purposes. The parser syntax should be what makes sense for hte user, and the
query structure constructed should be what makes sense for hte index, based on
the syntax used by the user.
If i have built an index consisting entirely of ngrams, the end user shouldn't
have to know that -- they shouldn't have to put every individual word in quotes
to force a PhraseQuery to be constructed out of the ngram tokenstream produced
by an individual word.
bq. Why isn't english treated this way too? I don't consider this bias towards
english "at all costs" including preventing languages such as Chinese from
working at all very fair, I think its a really ugly stance for Lucene to take.
I personally don't view it as an "english bias" ... to be it is a "backwards
compatibility bias"
I'm totally happy to make things onfigurable, but if two diametrically opposed
behaviors are both equally useful, and If there is a choice needs to be made
between leaving the default configuration the way the current hardcoded
behavior is, or make the default the exact opposite of what the current
hardcoded behavior is, it is then i would prefer to leave the default alone --
especially since this beahior has been around for so long, and many Analyzers
and TOkenFilters, have been written with this behavior specificly in mind
(several examples of this are in the Lucene code base -- and if we have them in
our own code, you can be sure they exist "in the wild" of client code that
would break if this behavior changes by default)
Once again: if this is a problem that can be solved "per instance" with token
attributes, then by all means let's make *all* of the TokenFIlters that come
"out of the box" implement this appropriately (english and non-english alike)
so that people who change the default settings on the queryparser get the
"correct" behavior regardless of langauge. but all other things being equal
lets keep the behavior working the way it has to avoid suprises.
bq. What are some real use-cases where this is "good"?
* WordDelimiterFilter ("wi-fi" is a pathalogicaly bad example for this issue
because as robert pointed out "wi" and "fi" don't tend to exist independently
in english, but people tend to get anoyed when "race-horse" matches all docs
containing "race" or "horse")
* single word to multiword synonym expansion/transformation (particularly
acronym expansion: GE => General Electric)
* Ngram indexing for fuzzy matching (if someone searches for the word
billionaire they're going to be surprised to get documents containing "lion")
> queryparser shouldn't generate phrasequeries based on term count
> ----------------------------------------------------------------
>
> Key: LUCENE-2458
> URL: https://issues.apache.org/jira/browse/LUCENE-2458
> Project: Lucene - Java
> Issue Type: Bug
> Components: QueryParser
> Reporter: Robert Muir
> Priority: Critical
>
> The current method in the queryparser to generate phrasequeries is wrong:
> The Query Syntax documentation
> (http://lucene.apache.org/java/3_0_1/queryparsersyntax.html) states:
> {noformat}
> A Phrase is a group of words surrounded by double quotes such as "hello
> dolly".
> {noformat}
> But as we know, this isn't actually true.
> Instead the terms are first divided on whitespace, then the analyzer term
> count is used as some sort of "heuristic" to determine if its a phrase query
> or not.
> This assumption is a disaster for languages that don't use whitespace
> separation: CJK, compounding European languages like German, Finnish, etc. It
> also
> makes it difficult for people to use n-gram analysis techniques. In these
> cases you get bad relevance (MAP improves nearly *10x* if you use a
> PositionFilter at query-time to "turn this off" for chinese).
> For even english, this undocumented behavior is bad. Perhaps in some cases
> its being abused as some heuristic to "second guess" the tokenizer and piece
> back things it shouldn't have split, but for large collections, doing things
> like generating phrasequeries because StandardTokenizer split a compound on a
> dash can cause serious performance problems. Instead people should analyze
> their text with the appropriate methods, and QueryParser should only generate
> phrase queries when the syntax asks for one.
> The PositionFilter in contrib can be seen as a workaround, but its pretty
> obscure and people are not familiar with it. The result is we have bad
> out-of-box behavior for many languages, and bad performance for others on
> some inputs.
> I propose instead that we change the grammar to actually look for double
> quotes to determine when to generate a phrase query, consistent with the
> documentation.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
