Interpreted GSQL Limitations

Currently, a query defined withCREATE DISTRIBUTED QUERYcannot be run in Interpreted Mode. However, interpreted queries can still run on a distributed graph with a regular, non-distributed execution model.

The list below includes additional limitations. These limitations are expected to be temporary. We are continuing to expand the capabilities of Interpreted Mode.

Unsupported Features in Interpreted Mode

  • Distributed Query Mode

  • Any operation involving files

    • FILE objects

    • PRINT TO_CSV

    • LoadAccum() function

    • SelectVertex() function

  • Statements

    • Exception statements:

      • RAISE statements

      • TRY...EXCEPTION statements

    • RETURN statement

      • Returning a vertex set variable

    • SELECT statement

      • ORDER BY clause

    • Assigning value to a global variable at the statement level

  • Attributes and accumulators

    • ArrayAccum

    • Accessing the previous value of a vertex-attached accumulator with the ' operator, e.g., src.@acc'

    • STRING COMPRESS as accumulator type

    • Accessing the attribute of a vertex variable that is not a vertex alias defined in a FROM clause

  • Functions and Operators

    • Aggregate functions: AVG(), MIN(), MAX(), COUNT(), ISEMPTY(), SUM()

    • neighbor(), neighborAttribute()

    • COALESCE()

    • IS NULL, IS NOT NULL

    • evaluate()

    • datetime_format()

    • parse_json_object(), parse_json_array()

    • User-Defined Functions

  • Data types

    • JSONOBJECT, JSONARRAY

    • BAG type parameters

  • Other

    • In INSERT: Optional vertex type information for from- and to- vertices when inserting edges

    • Using _ or ANY to specify any source vertex type in a FROM clause

    • Having the same name for global accumulators, global variables, and query parameters

      • For example, there cannot be a query parameter named param and a global accumulator named @@param within the same query