Why use different queries to query the same node, but the results are different?

Users Dgraph
1

What I want to do

I want to explore:
(1)、why use different queries to query the same node, but the results are different?
(2)、There are many data in my Dgraph. when i add eq space_id (eg:eq(space_id, 2323)) in filter , many datas will be Missed. How can i deal this question?

What I did

this is my first query

企业咚咚20210621162157
企业咚咚202106211621571571×338 49.6 KB

this is my second query

企业咚咚20210621162322
企业咚咚202106211623221468×310 51.3 KB

space_id in schema

企业咚咚20210621162608
企业咚咚202106211626081251×70 4.65 KB

Dgraph metadata

dgraph version

20.11

2

I am curious, what is the results of:

{
  queryA(func: type(JimdbShard)) @filter(eq(shard_id, 173305)) {
    uid
    dgraph.type
    shard_id
    space_id
  }
  queryB(func: type(JimdbShard)) @filter(eq(space_id, 2323)) {
    uid
    dgraph.type
    shard_id
    space_id
  }
  queryC(func: type(JimdbShard)) @filter(eq(shard_id, 173305) and eq(space_id, 2323)) {
    uid
    dgraph.type
    shard_id
    space_id
  }
  queryD(func: eq(shard_id, 173305)) {
    uid
    dgraph.type
    shard_id
    space_id
  }
  queryE(func: eq(space_id, 2323)) {
    uid
    dgraph.type
    shard_id
    space_id
  }
}
3

Sorry, I used upsert to fix the node. Now I can query out the data using eq(space_id, 2323).The result cannot be reproduced for now. But I feel that this kind of problem will still occur.

image
image1461×355 51.4 KB

4

What exactly did you do with an upsert to fix it?

5

I get the same problem!
‘space_id’: 5032,
‘shard_id’: 136416

result_test.txt (229.0 KB)

6

Use set in ratel to fix:
set {
space_id 2323
}

7

@amaster507 @MichelDiz Can you give me some advise？

8

No idea. Try to clean up the whole cluster(after a backup/export of course). Make sure your cluster has no issues like “cannot connect to alpha X or Y”. Or try to rebuild both indexes.

We need something solid to reproduce in order to help.