AI OnAI Off
Hi,
It was filed as a bug and has been fixed in upcoming version (possibly 8.9): http://world.episerver.com/support/Bug-list/bug/121067
A temporary fix for you is to change ecf_CatalogNode_GetAllChildEntries to this:
create procedure ecf_CatalogNode_GetAllChildEntries @catalogNodeIds udttCatalogNodeList readonly as begin with all_node_relations as ( select ParentNodeId, CatalogNodeId as ChildNodeId from CatalogNode where ParentNodeId > 0 union select ParentNodeId, ChildNodeId from CatalogNodeRelation ), hierarchy as ( select n.CatalogNodeId, '|' + CAST(n.CatalogNodeId as nvarchar(4000)) + '|' as CyclePrevention from @catalogNodeIds n union all select children.ChildNodeId as CatalogNodeId, parent.CyclePrevention + CAST(children.ChildNodeId as nvarchar(4000)) + '|' as CyclePrevention from hierarchy parent join all_node_relations children on parent.CatalogNodeId = children.ParentNodeId where CHARINDEX('|' + CAST(children.ChildNodeId as nvarchar(4000)) + '|', parent.CyclePrevention) = 0 ) select distinct ce.CatalogEntryId, ce.ApplicationId, ce.Code from CatalogEntry ce join NodeEntryRelation ner on ce.CatalogEntryId = ner.CatalogEntryId where ner.CatalogNodeId in (select CatalogNodeId from hierarchy) end
Regards.
/Q
Hi Quan,
Thanks, this temporary fix appears to have done the trick for now.
Jim
Hi,
EPiServer Commerce version 8.71
We're finding that adding new nodes to our EPiServer catalogue is very slow, and often times out, though all other catalogue modification operations seem fairly speedy. In particular, the method SaveCatalogNode in ICatalogSystem can take around 30 seconds to complete for each node being added.
I've run SQL Profiler to see what's going on, and the culprit appears to be this call:
Running this in SSMS tells me that this proc will retrieve ALL entries in our entire catalogue - but I have no idea what the SaveCatalogNode method might do with this information. Can anyone enlighten as to why our entire catalogue of entries (250,000+) has to be retrieved every time we want to add a new node? Any hints on how we might be able to speed this up?
Thanks,
Jim