Struct CombinedExpressionEvaluator 

pub struct CombinedExpressionEvaluator<'a> { /* private fields */ }

Evaluates expressions with combined schema (for JOINs)

Implementations

impl CombinedExpressionEvaluator<'_>

pub fn apply_affinity_for_comparison( &self, left_expr: &Expression, left_val: SqlValue, right_expr: &Expression, right_val: SqlValue, ) -> (SqlValue, SqlValue)

Apply SQLite affinity rules for comparisons.

When comparing a TEXT-affinity column to an INTEGER literal, SQLite:

1. Converts the INTEGER to TEXT
2. Performs string comparison

This method is public for use in GROUP BY expression evaluation.

pub fn apply_affinity_for_in_comparison( &self, left_expr: &Expression, left_val: SqlValue, right_expr: &Expression, right_val: SqlValue, ) -> (SqlValue, SqlValue)

Apply SQLite type affinity rules for IN expression comparisons.

IN expressions have different affinity rules than regular comparisons:

• For INTEGER columns, string literals are NOT coerced to integers
• For REAL columns, string literals ARE coerced to REAL
• For TEXT columns, numeric literals are converted to text

impl<'a> CombinedExpressionEvaluator<'a>

pub fn set_outer_rows(&mut self, outer_rows: &'a [Row])

Set all outer rows for outer-correlated aggregates (issue #4930).

When an aggregate function in a scalar subquery references only outer columns, it should aggregate over ALL outer rows, not just the current one.

pub fn clone_for_new_expression(&self) -> Self

Clone the evaluator for evaluating a different expression

Shares the subquery cache (safe because non-correlated subqueries produce the same results regardless of the current row) but creates a fresh CSE cache (necessary because CSE results depend on row values). Also shares correlation and hash caches (keyed by AST pointer, not row-dependent).