[openCypher tck case] empty list bug

[czpmango]

(czp@nebula) [nba]> yield [[2],[3]]
+------------+
| [[2],[3]]  |
+------------+
| [[2], [3]] |
+------------+
Got 1 rows (time spent 28917/41273 us)

Wed, 20 Jan 2021 14:36:56 CST

(czp@nebula) [nba]> yield [[2],[3],[]]
[ERROR (-7)]: SyntaxError: syntax error near `]]'

[jievince]

If we support empty list, maybe we should support empty set and empty map too, but:

empty list: []
empty set: {}
empty map: {}

empty set and empty map are in the same format so they're conflicted in the parser.
So we couldn't support empty set and empty map.
So should we support empty list?

[whitewum]

If we support empty list, maybe we should support empty set and empty map too, but:

empty list: []
empty set: {}
empty map: {}

empty set and empty map are in the same format so they're conflicted in the parser.
So we couldn't support empty set and empty map.
So should we support empty list?

where is the definition of set in opencypher tck?

[jievince]

If we support empty list, maybe we should support empty set and empty map too, but:

empty list: []
empty set: {}
empty map: {}

empty set and empty map are in the same format so they're conflicted in the parser.
So we couldn't support empty set and empty map.
So should we support empty list?

where is the definition of set in opencypher tck?

cypher doesn't support set. But we support set.

[whitewum]

If we support empty list, maybe we should support empty set and empty map too, but:

empty list: []
empty set: {}
empty map: {}

empty set and empty map are in the same format so they're conflicted in the parser.
So we couldn't support empty set and empty map.
So should we support empty list?

where is the definition of set in opencypher tck?

cypher doesn't support set. But we support set.

If you try to add a new concept, add some set tck cases like the map/list cases in opencypher tck feature files.

So, we can discuss whether the set behavior is proper and compatible with opencypher's list and map.

Both lists and maps have a lot of behavior definitions and specifications. So I don't think adding a new concept is as easy as we expect.

[jievince]

I agree, there are a lot of incompatible syntaxes

[czpmango]

IMO, set is distinct list.

[whitewum]

IMO, set is distinct list.

IMO， set in not necessary. Why, Cypher, has no concept set, in about ten years.

[czpmango]

IMO, set is distinct list.

IMO， set in not necessary. Why, Cypher, has no concept set, in about ten years.

ACK.