Struct InsertStatement 

pub struct InsertStatement {
    pub ctes: Vec<Cte>,
    pub table: String,
    pub alias: Option<String>,
    pub columns: Option<Vec<String>>,
    pub rows: Vec<Vec<Expr>>,
    pub select_source: Option<Box<SelectStatement>>,
    pub on_conflict: Option<OnConflictClause>,
    pub returning: Option<Vec<SelectItem>>,
    pub overriding: Overriding,
    pub mysql_ignore: bool,
}

Fields

ctes: Vec<Cte>

v7.37.43-T4.4 — leading WITH cte AS (…) clauses on a top- level INSERT (writable CTE outer body). Empty for a plain INSERT. PG semantics: each CTE materialises before the outer INSERT runs, sharing the same transaction.

table: String

alias: Option<String>

v7.39 (round 240) — INSERT INTO t AS alias: the alias the ON CONFLICT DO UPDATE expressions (and RETURNING) refer to the target row by. PG requires the AS keyword in this position.

columns: Option<Vec<String>>

Optional column list — INSERT INTO t (a, b) VALUES (...). When None, every tuple is positional and must match the table arity. When Some, the engine maps each tuple slot to the named column and fills the rest with NULL (must be nullable).

rows: Vec<Vec<Expr>>

One or more (expr, expr, ...) tuples — the multi-row VALUES form. v1.3+ accepts INSERT INTO t VALUES (a), (b). Empty when select_source is Some (the engine builds rows from the inner SELECT result set instead).

select_source: Option<Box<SelectStatement>>

v7.13.0 — INSERT INTO t [(cols)] SELECT … (mailrs round-5 G4). When present, rows is empty and the engine materialises the SELECT result, coerces each output tuple to the target column types, and inserts as a single batch.

on_conflict: Option<OnConflictClause>

v7.9.7 — ON CONFLICT (cols) DO { NOTHING | UPDATE SET … } upsert clause. None = legacy INSERT (conflict raises a DuplicateKey error). mailrs migration blocker #2.

returning: Option<Vec<SelectItem>>

v7.9.4 — RETURNING <projection>.

overriding: Overriding

v7.38 (read01) — OVERRIDING { SYSTEM | USER } VALUE clause between the column list and VALUES. Governs how explicitly-supplied values interact with GENERATED … AS IDENTITY columns:

• None — default. A GENERATED ALWAYS identity column rejects an explicit non-DEFAULT value; a BY DEFAULT one accepts it.
• System — override the ALWAYS restriction: the explicit value is used verbatim, as for a BY DEFAULT column.
• User — ignore any explicit value on a BY DEFAULT identity column and generate from the sequence instead (no effect on non-identity columns).

mysql_ignore: bool

v7.39 (round 434) — the statement was spelled INSERT IGNORE. Round 406 lowered that to ON CONFLICT DO NOTHING, which covers the key-conflict half of MySQL’s IGNORE. The other half is that IGNORE also downgrades per-VALUE errors to coercions (out-of-range clamps, over-long strings truncate, a non-numeric string becomes 0, a NULL into a NOT NULL column becomes the type’s default), and the engine cannot recover that intent from the conflict clause alone. A plain bool lands in this struct’s existing padding, so the AST does not grow — measured, per the round-305 / 413 nesting-stack lesson.

Trait Implementations

impl Clone for InsertStatement

fn clone(&self) -> InsertStatement

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for InsertStatement

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Display for InsertStatement

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl PartialEq for InsertStatement

fn eq(&self, other: &InsertStatement) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for InsertStatement