Tags: dgraph-io/dgraph
Tags
perf(core): Fix performance issue in type filter (#9065) Currently when we do queries like `func(uid: 0x1) @filter(type)`. We retrieve the entire type index. Sometimes, when the index is too big, fetching the index is quite slow. We realised that if we know we only want to check few `uids` are of the same, then we can just check those `uids` directly. Right now we are hard coding the number of `uids` threshold. This could be improved with a more statistical based model, where we figure out how many items does the type index have, how many we need to check.
Added similar_to in vector (#9062) This diff adds similar_to function to vectors, and adds the index to support it --------- Co-authored-by: Harshil Goel <harshil@dgraph.io>
fix(acl): allow data deletion for non-reserved predicates (#8937) when data (non-reserved predicates) is added on UIDs that belong to groot user or guardian group, it is allowed. But when the same data is deleted, that is not allowed. This PR allows deletion of non-reserved predicates on special UIDs. Closes: https://dgraph.atlassian.net/browse/DGRAPHCORE-355
PreviousNext