Rethink variable scope validation for path pattern

[yixinglu]

Please check the FAQ documentation before raising an issue

Describe the bug (required)

At present, the variable validation in path pattern only happens in match clause, but when the path pattern placed in the expression context, the variable's behavior will not be expected.

for example:

MATCH (v:player{name: 'Tim Duncan'})-[e:like*0..2]-(v2)
WHERE size([i in e WHERE (v)-[i]-(v2) | i])>1
RETURN count(*) AS cnt

//------

MATCH (v:player{name: 'Tim Duncan'})-[e:like*0..2]-(v2)-[i]-(v3)
WHERE size([i in e WHERE (v)-[i]-(v2) | i])>1
RETURN count(*) AS cnt

above tests are from the PR #5215 , and report the following error:

[ERROR (-1004)]: SyntaxError: syntax is ambiguous near `(v)-[i]-(v2)'

Your Environments (required)

• OS: uname -a
• Compiler: g++ --version or clang++ --version
• CPU: lscpu
• Commit id (e.g. a3ffc7d8)

How To Reproduce(required)

Steps to reproduce the behavior:

1. Step 1
2. Step 2
3. Step 3

Expected behavior

Additional context

[yixinglu]

@czpmango could you add more tests discussed offline for this topic?

[czpmango]

Variable shadowing error is unexpected:

(root@nebula) [nba]> match (v:player)-[e*2]->(n) where size([n in e where (v{name:"Tim Duncan"})-[n]-()])>3 return v
[ERROR (-1009)]: SemanticError: `n' is defined with type Node, but referenced with type Edge

---

Making an implicit join on the inner variable n and the outer variable n by mistake leads to an incorrect result. The expected row number is 4945(test in neo4j 4.2.1).

(root@nebula) [nba]>  match (v:player)-[e*2]->()-[n]-() where size([n in e where (v)-[n]-()])>0 return count(*)
+----------+
| count(*) |
+----------+
| 837      |
+----------+
Got 1 rows (time spent 132190/132459 us)

[xtcyclist]

Moved this issue to the 3.5 milestone. path pattern placed in the expression context does not seem to be an urgent issue for 3.4. 3.4 is mainly for stability.

[yixinglu]

related to #5235

[dutor]

Moved this issue to the 3.5 milestone. path pattern placed in the expression context does not seem to be an urgent issue for 3.4. 3.4 is mainly for stability.

Considering the amount of work, better to fix it in later version after v3.5.