Example of gtouping with empty groups in results:
Filed1 = rr_group, field2 = rr_updatedate
Problem is that I have tens of million groups in result and only several 
thousand with  "numFound" >2
   
"params":{
      "q":"*:* ",
      "group.sort":"rr_updatedate desc ",
      "group.limit":"-1",
      "fl":"rr_group,rr_adl,rr_createdate,rr_calctaskkey ",
      "group.offset":"2",
      "wt":"json",
      "group.field":"rr_group",
      "group":"true"}},
  "grouped":{
    "rr_group":{
      "matches":41475082,
      "groups":[{
          "groupValue":"164370:20200707:23:251",
          "doclist":{"numFound":1,"start":2,"docs":[]
          }},
        {
          "groupValue":"163942:20200708:22:251",
          "doclist":{"numFound":1,"start":2,"docs":[]
          }},
        {
          "groupValue":"163943:20200708:22:251",
          "doclist":{"numFound":1,"start":2,"docs":[]
          }},
        {
          "groupValue":"164355:20200708:22:251",
          "doclist":{"numFound":1,"start":2,"docs":[]

> -----Original Message-----
> From: Paras Lehana [mailto:paras.leh...@indiamart.com]
> Sent: Friday, February 14, 2020 3:37 PM
> To: solr-user@lucene.apache.org
> Subject: Re: Solr grouping with offset
> 
> It would be better if you give us an example.
> 
> On Fri, 14 Feb 2020 at 17:20, Vadim Ivanov
> <vadim.iva...@intourist.ru.invalid> wrote:
> 
> > Hello guys!
> > I need an advise. My task is to delete some documents in collection.
> > Del algorithm is following:
> > Group docs by field1  with sort by field2 and delete every 3 and
> > following occurrences in every group.
> > Unfortunately I didn't find easy way to do so.
> > Closest approach was to use group.offset = 2, but  result set is
> > polluted with empty groups with no documents (they have less then 3 docs
> in group).
> > May be I'm missing smth and there is way not to receive empty groups
> > in results?
> > Next approach was to use facet first with facet.mincount=3, then find
> > docs ids by every facet result  and then delete docs by id.
> > That way seems to me  too complicated for the task.
> > What's the best use case for the task?
> >
> 
> 
> --
> --
> Regards,
> 
> *Paras Lehana* [65871]
> Development Engineer, *Auto-Suggest*,
> IndiaMART InterMESH Ltd,
> 
> 11th Floor, Tower 2, Assotech Business Cresterra, Plot No. 22, Sector 135,
> Noida, Uttar Pradesh, India 201305
> 
> Mob.: +91-9560911996
> Work: 0120-4056700 | Extn:
> *11096*
> 
> --
> *
> *
> 
>  <https://www.facebook.com/IndiaMART/videos/578196442936091/>

Reply via email to