Hi Oleg,

do you mean that you are adding properties to the schema or that you are
adding property values to existing vertexes?
What I suggest you is:
- create schema with properties before the insert operations (this will
save space on the disk because property names are not stored in the record
as long as the property is defined in the schema)
- add vertex properties when you create the vertex, eg.
graph.addVertex("class:MyClass", "property1", "property1Val", "property2",
"property2Val");
- use Tx but commit every (say) 1000 operations
- use plocal instead of remove (if you can), it will increase the insert
performance a lot
- use MassiveInsert intent
http://www.orientechnologies.com/docs/last/orientdb.wiki/Performance-Tuning-Graph.html
- defining indexes before or after the insert will not make a big difference

Regards

Luigi


2015-01-30 11:13 GMT+01:00 Oleg Proudnikov <[email protected]>:

> Hi All,
>
> When I add a simple STRING property to a vertex class even without any
> constraints, it takes a significant amount of time. I do this after bulk
> graph load. I am wondering whether it is better to create properties at the
> time of schema creation, before the load. Would this save time?
>
> I also create indices after the bulk load. I wonder if creating indices
> beforehand would be more efficient. My goal is to reduce total time to
> build an indexed graph. I currently have 160M vertices with graph directory
> footprint of 40G, presumably mostly taken by the vertex properties.
>
> Thank you for any thoughts,
>
> Oleg
>
>  --
>
> ---
> 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.
>

-- 

--- 
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.

Reply via email to