If you have just 5-7 items, then an external file will work, as will the
join query. You'll need to handle the 'default' case with the join
query, that is, making sure you do <some query> OR <join query> so that
documents matching the join are boosted above those matching the main
query, rather than the join being a filter on the main query.
I can provide examples if needed.
Upayavira
On Mon, Sep 7, 2015, at 07:21 PM, Aman Tandon wrote:
> I am currently doing boosting for 5-7 things. will it work great with
> this
> too?
>
> With Regards
> Aman Tandon
>
> On Mon, Sep 7, 2015 at 11:42 PM, Upayavira <u...@odoko.co.uk> wrote:
>
> > External file field would work, but requires a full import of the
> > external file field every time you change a single entry, which is
> > pretty extreme.
> >
> > I've tested out "score joins" which seemed to perform very well and
> > achieved the same effect, but using another core, rather than an
> > external file.
> >
> > Thus:
> >
> > {!join score=max fromIndex=prices from=id to=id}{!boost b=price}*:*
> >
> > seemed to do the job of using the price as a boost. Of course you could
> > extend this like so:
> >
> > q={!join score=max fromIndex=prices from=id to=id}{!boost b=$b}*:*
> > b=sqrt(price)
> >
> > or such things to make the price a more reasonable value.
> >
> > Upayavira
> >
> > On Mon, Sep 7, 2015, at 06:21 PM, Aman Tandon wrote:
> > > Any suggestions?
> > >
> > > With Regards
> > > Aman Tandon
> > >
> > > On Mon, Sep 7, 2015 at 1:07 PM, Aman Tandon <amantandon...@gmail.com>
> > > wrote:
> > >
> > > > Hi Upayavira,
> > > >
> > > > Have you tried it?
> > > >
> > > >
> > > > No
> > > >
> > > > E.g. external file fields don't play nice with Solr Cloud
> > > >
> > > >
> > > > We are not using Solr Cloud.
> > > >
> > > >
> > > >> What are you using the external file for?
> > > >
> > > >
> > > > We are doing the boosting in the search result which are *having price
> > by
> > > > 1.2* & *country is India by 1.1*. We are doing by using the boosting
> > > > parameter in conjucation with query & map function e.g.
> > *&boost=map(query({!dismax
> > > > qf=hasPrice v='yes' pf=''},0),1,1,1,1)*
> > > >
> > > > This is being done with 5/6 parameters. And I am hoping it will
> > increase
> > > > query time. So I am planning to make the single score and populate it
> > in
> > > > external file field. And this might reduce some time.
> > > >
> > > > Just to mention we are doing incremental updates after every 10
> > minutes.
> > > >
> > > > With Regards
> > > > Aman Tandon
> > > >
> > > > On Mon, Sep 7, 2015 at 12:53 PM, Upayavira <u...@odoko.co.uk> wrote:
> > > >
> > > >> Have you tried it? I suspect your issue will be with the process of
> > > >> reloading the external file rather than consuming it once loaded.
> > > >>
> > > >> What are you using the external file for? There may be other ways
> > also.
> > > >> E.g. external file fields don't play nice with Solr Cloud.
> > > >>
> > > >> Upayavira
> > > >>
> > > >> On Mon, Sep 7, 2015, at 07:05 AM, Aman Tandon wrote:
> > > >> > Hi,
> > > >> >
> > > >> > How much ids information can I define in External File? Currently I
> > am
> > > >> > having the 100 Million records in my index.
> > > >> >
> > > >> > With Regards
> > > >> > Aman Tandon
> > > >>
> > > >
> > > >
> >
