Remove RDFGraph name syntactic sugar support from node variables (hashtag2782)

[vikramsubramanian]

We currently support using the RDFGraph name as a syntactic sugar for node variables. That is, suppose you have an RDFGraph named UniKG. Then you can do the following:

MATCH (a:UniKG)
RETURN a.iri, a.val

Output:
--------------------------------------------------------------
| a.iri                                           | a.val    |
--------------------------------------------------------------
|                                                 | Waterloo |
--------------------------------------------------------------
|                                                 | 150000   |
--------------------------------------------------------------
...
--------------------------------------------------------------
|                   |          |
--------------------------------------------------------------
|  |          |
--------------------------------------------------------------
...
--------------------------------------------------------------

I don't find this very valuable and this is extra code for us. We can just support the following multi-label querying to get the same output:

MATCH (a:UniKG_r:UniKG_l)
RETURN a.iri, a.val

We should still keep the UniKG rel group to simplify querying triples, since we already have a notion of rel groups. )

[mayil-ai[bot]]

Summary: Remove support for using RDFGraph name as syntactic sugar for node variables.

Possible Solution

Based on the provided information, the issue is about removing the feature that allows using RDFGraph names as syntactic sugar in node variable queries. To address this issue:

• Identify and remove the parser or interpreter code that handles the syntactic sugar for RDFGraph names. This could involve:

  • Removing or modifying functions that parse the syntactic sugar from the query strings.
  • Adjusting the grammar rules in the parser (e.g., in third_party/antlr4_cypher/cypher_parser.cpp) to no longer recognize the RDFGraph names as valid syntax for node variables.

• Update the query execution engine to ensure it no longer relies on the syntactic sugar feature. This may involve:

  • Refactoring the code that translates parsed queries into executable operations to exclude the handling of RDFGraph names as syntactic sugar.
  • Ensuring that multi-label querying is supported as an alternative method, which may require updating or adding functions similar to match_multi_label in the hypothetical code snippet.

• Update any unit tests or integration tests that rely on the syntactic sugar feature to use the alternative querying method (multi-label querying). This could involve:

  • Modifying test cases in files like test/c_api/value_test.cpp and tools/java_api/src/test/java/com/kuzudb/test/ValueTest.java to use the updated query syntax without RDFGraph names as syntactic sugar.
  • Ensuring that all tests pass with the updated query syntax and that the removal of the feature does not introduce regressions.

• Update documentation and examples to reflect the removal of the feature and guide users on how to use multi-label querying instead.

• Communicate the change to users, possibly through release notes or migration guides, to inform them of the removal of the syntactic sugar feature and the recommended alternative approach.

Code snippets to check

• third_party/antlr4_cypher/cypher_parser.cpp

  This file likely contains the grammar and parsing logic for Cypher queries, which would need to be updated to remove support for the RDFGraph name syntactic sugar.