Hi Savio,
Thanks for the reply. I have indexes and unique keys on my edges like so:
edges: loved UK:in_out, IDX:out, IDX:updatedAt
It's really when I combine the order by e.g.* order by in_watched.size() *with
the filter clause e.g.* inE('watched','loved','owns')[out=#12:25903] *does
it drop my 500ms
Leaving either out, and it is incredibly fast. I agree, that an index could
be the solution, as the explain shows:
@version
resultSize
documentReads
current
documentAnalyzedCompatibleClass
recordReads
fetchingFromTargetElapsed
fullySortedByIndex
orderByElapsed
evaluated
user
elapsed
resultType
indexIsUsedInOrderBy
0
2999
2999
#13:5189
<http://localhost:2480/studio/index.html#/database/test/browse/edit/13:5189>
2999
2999
650
false
3
2999
#5:0
<http://localhost:2480/studio/index.html#/database/test/browse/edit/5:0>
654.3758
collection
false
But I don't know how you index an edge count?
Look forward to hearing people's ideas! After 10hrs of experimenting, it's
getting very unproductive! :(
thanks
Sky
On Wednesday, 12 August 2015 13:41:16 UTC+1, SavioL wrote:
>
> Hi,
> certainly a way to increase performance is to index the fields on which you
> make the query (i do not know maybe you've already done).
> For stable which approach is better, I think the only way is to experiment
> and
> compare between them.
>
> Regards,
> Savio L.
>
--
---
You received this message because you are subscribed to the Google Groups
"OrientDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
