The URL path should be for one collection. This will be where the
collection is compiled and run. It has no effect on what is actually being
searched. That is specified in the expression themselves.
Joel Bernstein
http://joelsolr.blogspot.com/
On Wed, Jan 20, 2021 at 1:34 PM ufuk yılmaz
wrote:
> Do collection names in request url affect how the query works in any way?
>
> A streaming expression is sent to http://mySolrHost/solr/col1,col2/stream
> (notice multiple collections in url)
>
> Col1 has 2 shards, each have 3 replicas.
> * Shard1 has replicas on nodes A, B, C
> * Shard2 has replicas on D,E,F
>
> Col2 has 2 shards, each have 3 replicas. Its shards have the same
> configuration as Col1.
>
>
> Lets say we have a simple search expression:
> search(
> "colA,colB",
> q="*:*",
> qt="/export",
> fl="fl1,fl2",
> sort="id asc"
> )
>
> Collection names in search expression denotes which collections should be
> searched, so we can’t change them. But what would change if we sent the
> query to
> http://mySolrHost/solr/someOtherCollection/stream
>
> and someOtherCollection has 1 shard and 6 replicas in nodes A,B,C,D,E,F ?
>
> I read about worker collections a bit, but as long as I don’t explicitly
> use parallel streams, what is the difference?
>
>
>
> Sent from Mail for Windows 10
>
>
