Module df_expr 

Translation layer from Cypher expressions to DataFusion expressions.

This module provides cypher_expr_to_df which converts Cypher AST expressions into DataFusion physical expressions suitable for use in DataFusion execution plans.

Property Naming Convention

Properties are materialized as columns with the naming convention {variable}.{property}. For example, n.age becomes column "n.age".

Supported Expressions

• Identifiers and property access
• Literal values (numbers, strings, booleans, null)
• Binary operators (comparison, arithmetic, boolean)
• Unary operators (NOT, negation)
• IS NULL / IS NOT NULL
• String operations (CONTAINS, STARTS WITH, ENDS WITH)
• IN list checks
• CASE expressions

Unsupported Expressions

Some Cypher expressions require custom handling and are not yet supported:

• List comprehensions
• Reduce expressions
• Subqueries (EXISTS, scalar subqueries)
• Approximate equality (~=) for vectors

Structs

TranslationContext

Context for expression translation.

Enums

VariableKind

Entity kind of a variable in the physical query context.

Functions

apply_type_coercion

Apply type coercion to a DataFusion expression.

cast_expr

Create a cast expression to the specified data type.

collect_properties

Collect all property accesses from an expression tree.

comparison_udf_name

Map a comparison operator to its _cypher_* UDF name.

cypher_expr_to_df

Convert a Cypher expression to a DataFusion expression.

dummy_udf_expr

Helper to create a DummyUdf wrapped in a ScalarFunction expression.

extract_datetime_nanos

Extract the nanos_since_epoch field from a DateTime struct expression.

extract_time_nanos

Extract the UTC-normalized time in nanoseconds from a Time struct expression.

list_to_large_binary_expr

Wrap a List<T> or LargeList<T> expression as a LargeBinary CypherValue.

normalize_datetime_str

Insert :00 seconds into a datetime string like 2021-06-01T00:00Z that has only HH:MM after the T separator (no seconds component).

scalar_to_large_binary_expr

Wrap a native scalar expression (Int64, Float64, Utf8, Boolean, etc.) in the _cypher_scalar_to_cv UDF so it becomes CypherValue-encoded LargeBinary. Used to normalize mixed-type coalesce arguments.

struct_getfield

Extract a named field from a struct expression using DataFusion’s get_field function.

wider_numeric_type

Returns the wider of two numeric DataTypes for type coercion.