LIMIT <list> in GO always return the same result and can be stuck in loops

[randomJoe211]

Nebula Graph version: 2021.10.18-nightly

I have 5 vertices: "1" to "5". The edges are as follows:

insert edge follow(degree) values "1"->"2":(95);
insert edge follow(degree) values "2"->"3":(95);
insert edge follow(degree) values "3"->"1":(95);
insert edge follow(degree) values "2"->"4":(95);
insert edge follow(degree) values "4"->"5":(90);
insert edge follow(degree) values "3"->"4":(90);

If I run GO 3 STEPS FROM "1" OVER follow LIMIT [1,1,1], I always get the same result "1". This test showed that:

1. GO... LIMIT <list> always walks through the same path, probably depending on the dictionary order or the VIDs.
2. When there is a loop, it can be stuck forever.

Is it designed to be like this?

The test nGQL are as follows.

(root@nebula) [basketballplayer]> create space test(partition_num=10,replica_factor=1,vid_type=fixed_string(32));
Execution succeeded (time spent 1670/2211 us)

Fri, 22 Oct 2021 06:09:45 UTC

(root@nebula) [basketballplayer]> use test
[ERROR (-1005)]: SpaceNotFound:

Fri, 22 Oct 2021 06:09:48 UTC

(root@nebula) [basketballplayer]> use test
Execution succeeded (time spent 511/760 us)

Fri, 22 Oct 2021 06:09:51 UTC

(root@nebula) [test]> create tag player(name string,age int);
Execution succeeded (time spent 1184/1583 us)

Fri, 22 Oct 2021 06:10:46 UTC

(root@nebula) [test]> create tag team(name string);
Execution succeeded (time spent 783/1154 us)

Fri, 22 Oct 2021 06:10:48 UTC

(root@nebula) [test]> create edge follow(degree int);
Execution succeeded (time spent 915/1181 us)

Fri, 22 Oct 2021 06:11:13 UTC

(root@nebula) [test]> insert vertex player(name,age) values "1":("Tim Duncan", 42);
Execution succeeded (time spent 1436/1942 us)

Fri, 22 Oct 2021 06:11:46 UTC

(root@nebula) [test]> insert vertex player(name,age) values "2":("Tony Parker", 36);
Execution succeeded (time spent 647/807 us)

Fri, 22 Oct 2021 06:11:46 UTC

(root@nebula) [test]> insert vertex player(name,age) values "3":("LaMarcus Aldridge", 33);
Execution succeeded (time spent 421/530 us)

Fri, 22 Oct 2021 06:11:46 UTC

(root@nebula) [test]> insert vertex player(name,age) values "4":("Rudy Gay", 32);
Execution succeeded (time spent 402/508 us)

Fri, 22 Oct 2021 06:11:46 UTC

(root@nebula) [test]> insert vertex player(name,age) values "5":("Marco Belinelli", 32);
Execution succeeded (time spent 811/1215 us)

Fri, 22 Oct 2021 06:11:48 UTC

(root@nebula) [test]> insert edge follow(degree) values "1"->"2":(95);
Execution succeeded (time spent 1252/1635 us)

Fri, 22 Oct 2021 06:13:10 UTC

(root@nebula) [test]> insert edge follow(degree) values "2"->"3":(95);
Execution succeeded (time spent 613/755 us)

Fri, 22 Oct 2021 06:13:10 UTC

(root@nebula) [test]> insert edge follow(degree) values "3"->"1":(95);
Execution succeeded (time spent 480/599 us)

Fri, 22 Oct 2021 06:13:10 UTC

(root@nebula) [test]> insert edge follow(degree) values "2"->"4":(95);
Execution succeeded (time spent 485/623 us)

Fri, 22 Oct 2021 06:13:10 UTC

(root@nebula) [test]> insert edge follow(degree) values "4"->"5":(90);
Execution succeeded (time spent 401/511 us)

Fri, 22 Oct 2021 06:13:10 UTC

(root@nebula) [test]> insert edge follow(degree) values "3"->"4":(90);
Execution succeeded (time spent 607/812 us)

Fri, 22 Oct 2021 06:13:11 UTC

(root@nebula) [test]> GO FROM "1" OVER follow LIMIT [1]
+-------------+
| follow._dst |
+-------------+
| "2"         |
+-------------+
Got 1 rows (time spent 1665/2244 us)

Fri, 22 Oct 2021 06:14:37 UTC

(root@nebula) [test]> GO FROM "1" OVER follow LIMIT [1]
+-------------+
| follow._dst |
+-------------+
| "2"         |
+-------------+
Got 1 rows (time spent 1058/1560 us)

Fri, 22 Oct 2021 06:14:41 UTC

(root@nebula) [test]> GO FROM "1" OVER follow LIMIT [1]
+-------------+
| follow._dst |
+-------------+
| "2"         |
+-------------+
Got 1 rows (time spent 814/1121 us)

Fri, 22 Oct 2021 06:14:42 UTC

(root@nebula) [test]> GO 3 STEPS FROM "1" OVER follow LIMIT [1,1,1]
+-------------+
| follow._dst |
+-------------+
| "1"         |
+-------------+
Got 1 rows (time spent 2137/2465 us)

Fri, 22 Oct 2021 06:15:21 UTC

(root@nebula) [test]> GO 3 STEPS FROM "1" OVER follow LIMIT [1,1,1]
+-------------+
| follow._dst |
+-------------+
| "1"         |
+-------------+
Got 1 rows (time spent 2023/2307 us)

Fri, 22 Oct 2021 06:15:23 UTC

(root@nebula) [test]> GO 3 STEPS FROM "1" OVER follow LIMIT [1,1,1]
+-------------+
| follow._dst |
+-------------+
| "1"         |
+-------------+
Got 1 rows (time spent 1783/1960 us)

Fri, 22 Oct 2021 06:15:24 UTC

(root@nebula) [test]> GO 3 STEPS FROM "1" OVER follow LIMIT [1,1,1]
+-------------+
| follow._dst |
+-------------+
| "1"         |
+-------------+
Got 1 rows (time spent 1485/1682 us)

Fri, 22 Oct 2021 06:15:25 UTC

(root@nebula) [test]> GO 3 STEPS FROM "1" OVER follow LIMIT [1,1,1]
+-------------+
| follow._dst |
+-------------+
| "1"         |
+-------------+
Got 1 rows (time spent 1577/1796 us)

Fri, 22 Oct 2021 06:15:26 UTC

(root@nebula) [test]> GO 3 STEPS FROM "1" OVER follow LIMIT [1,1,1]
+-------------+
| follow._dst |
+-------------+
| "1"         |
+-------------+
Got 1 rows (time spent 1760/2128 us)

[Shylock-Hg]

1. What's the problem?
2. The detail case?

[jievince]

When there is a loop, it can be stuck forever.

How does it be stuck? Can you explain it in detail?

[randomJoe211]

When there is a loop, it can be stuck forever.

How does it be stuck? Can you explain it in detail?

@jievince Suppose we have 1->2->3->1, and run GO FROM "1" OVER * LIMIT [1,1,1], the result will always be 1 itself.
If we have 1->2 and 2->1, and 1-> a number of other vertices, the first 2 rows returned are always 1 and 2, no matter how we set the list. I just think this may not be what the users would want.