[
https://issues.apache.org/jira/browse/SOLR-15407?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17344695#comment-17344695
]
Alessandro Benedetti edited comment on SOLR-15407 at 5/14/21, 4:22 PM:
-----------------------------------------------------------------------
I opened an initial Pull Request to mainly discuss what to do with queries
containing numerical terms, numerical query fields and sow=false
There were numerous tests in
org.apache.solr.search.TestExtendedDismaxParser#testFocusQueryParser for those
use cases.
e.g.
query=terminator 100
queryFields= field_text_general integer_i
sow=false
What do we want as a final query?
*option1* (we are consistent with sow=false and not split on whitespace happens
happens also for the numerical field)
field_text_general:(terminator 100)
<integer_i:(terminator 100) -> this bit in disjunction disappears because query
value not a number)>
OR
*option2* (we bring inconsistency, so for numerical fields we basically ignore
sow and just split on whitespace and build boolean queries)
field_text_general:(terminator 100) | integer_i:(100)
[~mdrob][~munendrasn][~romseygeek][~dsmiley][~sarowe] what do you think?
Once agreed I'll clean up the PR and add additional tests/changes if needed
was (Author: alessandro.benedetti):
I opened an initial Pull Request to mainly discuss what to do with queries
containing numerical terms, numerical query fields and sow=false
There were numerous tests in
org.apache.solr.search.TestExtendedDismaxParser#testFocusQueryParser for those
use cases.
e.g.
query=terminator 100
queryFields= field_text_general integer_i
sow=false
What do we want as a final query?
*option1* (we are consistent with sow=false and not split on whitespace happens
happens also for the numerical field)
field_text_general:(terminator 100)
| <integer_i:(terminator 100) -> this disappears because query value not a
number)>
OR
*option2* (we bring inconsistency, so for numerical fields we basically ignore
sow and just split on whitespace and build boolean queries)
field_text_general:(terminator 100) | integer_i:(100)
[~mdrob][~munendrasn][~romseygeek][~dsmiley][~sarowe] what do you think?
Once agreed I'll clean up the PR and add additional tests/changes if needed
> eDismax sow=false doesn't work with string field types
> ------------------------------------------------------
>
> Key: SOLR-15407
> URL: https://issues.apache.org/jira/browse/SOLR-15407
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Components: query parsers
> Affects Versions: 8.8.2
> Reporter: Alessandro Benedetti
> Priority: Major
>
> Currently, the sow=false should not tokenize the input user query text and
> delegate to each field for query time text analysis.
> But what happens if one of the queries involved is not analyzed?
> For example, because it is a string field type?
> Terms are split and the query generated is broken:
> {code:java}
> assertU(adoc("id", "75", "trait_ss", "multi term"));
> public void testSplitOnWhitespace_stringField_shouldBuildSingleClause()
> throws Exception
> {
> assertJQ(req("qf", "trait_ss", "defType", "edismax", "q", "multi
> term", "sow", "false"),
> "/response/numFound==1", "/response/docs/[0]/id=='75'");
> String parsedquery;
> parsedquery = getParsedQuery(
> req("qf", "trait_ss", "q", "multi term", "defType", "edismax",
> "sow", "false", "debugQuery", "true"));
> assertThat(parsedquery, anyOf(containsString("((trait_ss:multi
> term))")));
> }
> {code}
> This test would be currently broken.
> The current parsed query is wrongly:
> (trait_ss:multi trait_ss:term)
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
