That is actually the current behavior of the search expression. The initial use cases from Streaming Expressions revolved around joins and rollups which really require the entire result set. So the search expression just merged the results from the shards and let the wrapping expression deal with the results. Things have evolved quite a bit since then and having the search expression respect the rows parameter is something that I've been meaning to add. Feel free to create a ticket for this.
Joel Bernstein http://joelsolr.blogspot.com/ On Thu, Jun 21, 2018 at 1:35 PM, Alfonso Muñoz-Pomer Fuentes < amu...@ebi.ac.uk> wrote: > I’m having a weird issue with the search streaming expressions and I’d > like to share it before opening a ticket in Jira, just in case I’m missing > something obvious. > > I’m currently on Solr 7.1 and I have a collection named bioentities split > into two shards and no replicas. Whenever I run a query such as this: > search( > bioentities, > q="*:*", > fl="bioentity_identifier,property_value,property_name", > sort="bioentity_identifier asc") > > I’m getting 20 documents. If I add e.g. rows=4 I get 8 results, and so on. > > I have the same collection in another SolrCloud cluster, split into three > shards and running the same queries I get 30 and 12 results, respectively. > So it seems that the seach expression distributes the query between shards > and then aggregates the results. Is this the expected behaviour? > > Thanks in advance. > > -- > Alfonso Muñoz-Pomer Fuentes > Senior Lead Software Engineer @ Expression Atlas Team > European Bioinformatics Institute (EMBL-EBI) > European Molecular Biology Laboratory > Tel:+ 44 (0) 1223 49 2633 > Skype: amunozpomer > >
