Hi,

I personally like the symmetry between edge and vertex labels. If you want to 
"change a label" for an edge, you have to delete the edge and create a new one 
with the label you want. Likewise for vertices. With this, label-methods fall 
nicely from Element. Finally, realize that there are also VertexProperty labels 
and would those support mutation ?… eek.

I have no real skin in the game, so….

Marko.

http://markorodriguez.com

On Nov 30, 2015, at 1:15 PM, Stephen Mallette <[email protected]> wrote:

> I'm only at the point of figuring out if there are enough graph providers
> who can implement single label mutability to warrant the discussion of
> whether or not we should have it in TinkerPop.  Thanks for weighing in from
> sqlg perspective.
> 
> On Mon, Nov 30, 2015 at 2:47 PM, pieter <[email protected]> wrote:
> 
>> Sqlg does not support changing a vertex label.
>> In our case however it could be as trivial as a ALTER TABLE command.
>> 
>> Neo4j supports many labels per node/element which is significantly
>> different to TinkerPop's label semantic.
>> Would you propose that also or just editing the one label?
>> 
>> Thanks
>> Pieter
>> 
>> 
>> 
>> 
>> On 30/11/2015 21:29, Stephen Mallette wrote:
>>> It occurred to me today while responding to a post regarding Titan that
>> TP
>>> 3 APIs assume that vertex labels are immutable. Titan is happy with that
>>> because it doesn't allow such things, but it made me wonder about other
>>> providers.
>>> 
>>> A quick bit of research showed that Neo4j could support such a thing:
>>> 
>>> http://neo4j.com/docs/stable/query-set.html#set-set-a-label-on-a-node
>>> 
>>> I was wondering if other providers could comment on their ability to
>>> support modifying a vertex label after the vertex is created.  If we find
>>> that it is such that many providers could support it then maybe the TP3
>> API
>>> should support that.
>>> 
>> 
>> 

Reply via email to