Docker _upload dir cleanup uses inefficient query

Docker cleanup is triggered after manifest deployment and uses the following query:

select distinct  n.repo as itemRepo,n.node_id as itemId,n.node_path as itemPath,n.node_name as itemName,n.node_type as itemType from  nodes n where ( n.repo = 'jfrog-docker' and n.depth >= 4 and( n.node_path like 'x/y/_uploads%' escape '^' or n.node_path = 'x/y/_uploads')and n.node_type = 1);

While Artifactory has the {{_upload }}dir location and and since all the files are written directly to the root of the {{_upload }}dir, the >= can be replaces with = and the {{LIKE % }}is unnecessary.

On medium/large env with dozens/hundreds millions of artifacts, it can reach to seconds per query vs millis with an optimized query