N̶o̶ ̶i̶d̶e̶a̶.̶ ̶W̶o̶r̶k̶s̶ ̶f̶o̶r̶ ̶m̶e̶.̶ ̶D̶o̶ ̶y̶o̶u̶ ̶s̶e̶e̶ ̶a̶n̶y̶ ̶e̶r̶r̶o̶r̶s̶?̶ ̶N̶o̶ ̶e̶r̶r̶o̶r̶s̶ ̶u̶s̶u̶a̶l̶l̶y̶ ̶m̶e̶a̶n̶ ̶s̶o̶m̶e̶ ̶v̶a̶r̶i̶a̶b̶l̶e̶s̶ ̶h̶a̶v̶e̶ ̶w̶r̶o̶n̶g̶ v̶a̶l̶u̶e̶s̶.̶ ̶ ̶M̶a̶k̶e̶ ̶s̶u̶r̶e̶ ̶a̶l̶l̶ ̶y̶o̶u̶r̶ ̶v̶a̶r̶i̶a̶b̶l̶e̶s̶ ̶a̶r̶e̶ ̶c̶o̶r̶r̶e̶c̶t̶,̶ ̶l̶i̶k̶e̶ ̶t̶h̶e̶ ̶u̶i̶d̶ ̶o̶n̶e̶s̶.̶ ̶A̶l̶s̶o̶ ̶m̶a̶k̶e̶ ̶s̶u̶r̶e̶ ̶y̶o̶u̶ ̶u̶s̶e̶ ̶n̶e̶w̶e̶s̶t̶ ̶d̶g̶r̶a̶p̶h̶ ̶v̶e̶r̶s̶i̶o̶n̶.̶
ayy… Sorry. Just noticed my mistake. Problem with filtering on uid is that uid always exist. It should be filtered not on uid but a different predicate. So in query when querying eq(myID, "123") will work fine. Or if we need uid, then add additional filter or directive to validate the node exists. @cascade or @filter(has(dgraph.type)) should be enough. I fixed and edited the query above.