The blog you're referring to is explaining the query time join
implementation that isn't using ordinal map, which is the first method in
the JoinUtil class (with the ordinal map as parameter).
If you want to use the method that uses global ordinals you can use this
snippet to create an OrdinalMap
Hi Alex,
I never took joining based on multiple indices into account when creating
the global ordinal join.
I like your idea of using a multi reader to make the join think it is
running on one index. But I don't know if there are edge cases where this
might not work.
Martijn
On 21 July 2015 at 2
ke that my problem isn't so much with the
> join utility, but more with my query parser plugging class. Is there
> something that it missing in the above link example that I need to also add
> to mine to ensure that queries are applied pre join? Thanks.
>
> -Shane
>
>
&
Hi,
If you specify the FilteredQuery via the 'fromQuery' parameter, then they
should be executed before the the join operation is performed.
What makes you think that the FilteredQuery is applied post join?
Martijn
On 1 August 2013 22:33, Shane Strasser wrote:
> Hey All,
>
> We're currently l
I don't recall that the RawTermFilter was required. The following code
should also work in 4.x:
Filter parentsFilter = new CachingWrapperFilter(new
QueryWrapperFilter(new TermQuery(new Term("type", "T1";
Martijn
On 16 January 2013 16:51, kiwi clive wrote:
> Hi Guys,
>
> Apologies if this has
more expensive (execution wise).
I hope this gives you some pointers to start with.
Martijn
On 11 December 2012 02:55, dizh wrote:
> Thanks to Martijn v Groningen.
>
> but , Is there anyone who has implemented this feature? just like SQL :
> select a1,a2,sum(a3) group by a1,a2;
>
&g
Hi Dish,
Grouping on multiple fields or a field that has multiple tokens (or
values) per document hasn't been implemented yet.
Martijn
On 10 December 2012 07:12, dizh wrote:
> Hi All:
> I want to ask how to do "Group by on multi fields ".
> The Lucene JavaDOC only gives a demo about how to grou
A top level document ID can change over time. For that reason you
shouldn't rely on it. However if you know your index is stable or you
keep track when a merge happes, you can use the
IndexWriter#tryDeleteDocument method to delete a document by Lucene
id. Deleting a document via a IndexReader is no
Lucene isn't a relation database. However there are some methods to
have a relational like search:
1) Index time joins:
http://blog.mikemccandless.com/2012/01/searching-relational-content-with.html
2) Query time joins:
http://www.searchworkings.org/blog/-/blogs/query-time-joining-in-lucene
Martijn
Hi Jayendra,
This isn't supported yet. You could implement this by creating a
custom Lucene collector.
This collector could count the unique hits inside a block of docs per
unique facet field value. The
unique facet values could be retrieved from Lucene's FieldCache or doc
values (if you can use L
If the time span or website (I assume you mean domain name) is a field
in your index then you can use result grouping.
Result grouping has impact on your query time and if you have a lot of
data you need to divide your data across multiple
indices and use distributed result grouping.
Martijn
On 2
Hi Christoph,
You can open an issue for this. I think we can use the child score as
an explanation of why a parent doc is scored the way it is.
Martijn
On 25 May 2012 13:20, Christoph Kaser wrote:
> Hello all,
>
> I try to calculate score explanations for a query that contains a
> ToParentBlock
It is only possible to join on one side at the same time.
You mean something like this:
Query fromQuery = JoinUtil.createJoinQuery("from", "to", actualQuery,
indexSearcher);
Query toQuery = JoinUtil.createJoinQuery("to", "from", actualQuery,
indexSearcher);
And then use a boolean query to combine
Some extra notes:
1) he implementation in 3.6 is about 3 times slower. I noticed this during
some tests that I ran on my test data set.
The 3.6 implementation is definitely usable. Also on larger indices. I
believe the average query time was around 350 ms on an index containing 10M
documents.
2) Th
Hi,
There is an example usage in the package.html.
Something to keep in mind is that this search time join is different than
block join query. The JoinUtil can be used to do join filtering and the
block join is more meant for parent / child search.
Martijn
On 23 March 2012 11:58, Li Li wrote:
http://nlp.stanford.edu/IR-book/information-retrieval-book.html gives
a good introduction what happens under the hood of a search engine and
you can download it for free. It does not explain Lucene directly, but
a lot of IR algorithms that are used in Lucene (and any other search
engine) are explai
16 matches
Mail list logo
