ent icon indicating copy to clipboard operation
ent copied to clipboard

ability to add cache on top of queries

Open lolopinto opened this issue 1 year ago • 0 comments

  • ent implementation should be storage agnostic. it should be a consistent API that clients can implement
  • will come with 3 implementations out of the box
    • memory [redundant with dataloader but worth just having as a simple test bed]
    • redis
    • memcache
  • will support following query types
    • query by id
    • query by edge type
    • ability to cache custom queries e.g. Foo.loadCustom(vc, clause) like what we already do with dataloader. probably don't want this by default since this this could affect many rows and if this is done by default and cleared all the time, the cache could become useless
  • automatically clearing the cache for known writes e.g. when object or edge is written
  • easy ability to to control clearing the cache when other things happen e.g. how to know when custom queries need to be cleared

don't have a use case for this yet but writing it down so that when use cases come, can reference it...

lolopinto avatar Jul 23 '23 23:07 lolopinto