Hi, Right now, there is only one nodetype index. So, if you add a nodetype / mixin to that index (as you know the lists of nodetypes / mixins is a multi-valued property), then you need to reindex that index. Which needs to read all the nodes.
The alternative would be to have multiple nodetype indexes. A patch for that is welcome! If you have that, then instead of changing the nodetype index, you create a new index. This also needs to read all the nodes. So, that would be more convenient (even though in both cases indexing a new nodetype takes about the same time). > Is this a design choice to only allow one nodetype index? Not a design choice, just how it is implemented right now. > I have no way to make it contained in the project itself to make an index > based on it's mixin type. There is a way, you need some code to extend the existing nodetype index, and then reindex. Regards, Thomas On 29.06.17, 16:42, "Roy Teeuwen" <r...@teeuwen.be> wrote: Hey all, Some time ago I asked about creating an oak index based on the node type (primary type or mixin type), after which I was pointed to the nodetype index. I have to say though that there is a serious drawback to this index: I have two separate projects, both having some code based on a mixinType. But seeing as there can only be one nodetype index, I have no way to make it contained in the project itself to make an index based on it's mixin type. Is this a design choice to only allow one nodetype index? Is there a workaround for this? Thanks! Roy