Hi,
I have a field which has data like this:
<letters>
<letters> <numbers>
<letters> <numbers> <letters> <numbers>
Where <letters> can have from 1 to 10 letters strings and <number> can have up
to 4 digits.
It is defined like this:
<field name="myField" type="myFieldType" indexed="true" stored="true"
multiValued="true" />
<fieldType name="myFieldType" class="solr.TextField"
positionIncrementGap="100">
<analyzer>
<tokenizer class="solr.KeywordTokenizerFactory" />
<filter class="solr.LowerCaseFilterFactory" />
</analyzer>
</fieldType>
When the user enters foo, i search for foo directly or something that starts
with "foo ".
I don't want to find "fool" or "foop" or anything like this.
I also allow users to enter terms that they don't want to find.
So a query for "foo NOT foo 123" is converted to this:
parsedquery_toString: "+(+(myField:foo myField:foo *) +(-myField:foo 123
-myField:foo 123 * +*:*))",
My problem is that this finds more entries then just foo, which converts to
this:
parsedquery_toString: "+(myField:foo myField:foo *)",
I have read a bit about the internal solr logic, using MUST, SHOULD and
MUST_NOT, but still don't understand.
When I look at parsedquery_toString: "+(+(myField:foo myField:foo *)
+(-myField:foo 123 -myField:foo 123 * +*:*))",
then I see two criteria A and B and both MUST be satisfied.
Criteria A is the same as parsedquery_toString: "+(myField:foo myField:foo *)",
so the number of results MUST be identical here.
Since the final results must match both A and B, the number must be equal or
lower than just A, right?
Where do I think wrong?
Thanks,
Kai
