V21.03: After pagination+cascade change, queries are too slow to finish

Issues Dgraph
,
13

sure:

with has and uid_in() (29.05s, 2896522 uids) 

  "extensions": {
    "server_latency": {
      "parsing_ns": 400591,
      "processing_ns": 29048607400,
      "encoding_ns": 474601,
      "total_ns": 29049622369
    },
    "txn": {
      "start_ts": 2995240
    },
    "metrics": {
      "num_uids": {
        "": 365468,
        "_total": 2896522,
        "qa.has_indicator": 6000,
        "qa.has_indicator.next": 156000,
        "qa.has_object": 106,
        "qa.has_object.next": 6000,
        "qa.has_timerange": 324002,
        "qa.name": 162002,
        "qa.timerange_end": 695471,
        "qa.timerange_start": 695471,
        "qa.type": 162000,
        "uid": 324002
      }
    }
  }

but in general the uid_in() function seems not useful to me to filter on my has_timerange edge - note my earlier example that expressed that. I would have to query the universe of timeranges, which will be billions, and use that as the uid variable argument: uid_in(<qa.has_timerange>,uid(allvalidtimeranges)) Maybe there is a way to use it for the other edges I want to filter on but on in the main path of the query, but the timeranges one especially, it seems like a no-go.

1 Like