Struct SecondaryIndex 

pub struct SecondaryIndex {
    pub name: String,
    pub table_name: String,
    pub column_name: String,
    pub is_unique: bool,
    pub origin: IndexOrigin,
    pub entries: IndexEntries,
}

One secondary index on a single column. Multi-column composite indexes are on the longer-term list; the column_name field stays singular for now.

Fields

name: String

Catalog name. For auto indexes: sqlrite_autoindex_<table>_<col>. For explicit indexes: the user-supplied identifier from CREATE INDEX.

table_name: String

column_name: String

is_unique: bool

origin: IndexOrigin

entries: IndexEntries

Implementations

impl SecondaryIndex

pub fn new( name: String, table_name: String, column_name: String, datatype: &DataType, is_unique: bool, origin: IndexOrigin, ) -> Result<Self>

Builds an empty index over a column of the given datatype. Returns an error for unsupported datatypes (Real, Bool, None, Invalid).

pub fn synthesized_sql(&self) -> String

Synthesizes a CREATE INDEX statement for sqlrite_master.sql. For auto indexes this is a synthetic form; for explicit indexes the caller can override with the original user text if it has been preserved. Used by the persistence path in Phase 3e.4.

pub fn auto_name(table_name: &str, column_name: &str) -> String

Standard name for the auto-generated index of a UNIQUE/PK column. Uniform across save/open so indexes persist under a stable name.

pub fn would_violate_unique(&self, value: &Value) -> bool

Returns true iff inserting value would violate the UNIQUE constraint — i.e., the index already has an entry for this value and self.is_unique is set. Null values are never indexed and never conflict.

pub fn insert(&mut self, value: &Value, rowid: i64) -> Result<()>

Adds a (value, rowid) entry. Null values are ignored (NULL in an indexed column stays out of the index). Type mismatches — e.g. calling this with a Text value against an Integer index — return an error rather than silently skipping.

pub fn remove(&mut self, value: &Value, rowid: i64)

Removes a (value, rowid) entry. If the value has other rowids attached they remain; if this was the last rowid, the value key is pruned. A no-op if the entry isn’t present (simpler than failing — UPDATE paths rely on this).

pub fn lookup(&self, value: &Value) -> Vec<i64>

Returns every rowid currently associated with value. For a unique index this is at most one; for a non-unique index it can be many. Empty Vec if the value isn’t present.

pub fn iter_entries(&self) -> Box<dyn Iterator<Item = (Value, i64)> + '_>

Iterates every (value, rowid) pair in ascending-value order. The rowids for a given value come out in insertion order, which happens to match ascending rowid order in practice because rows are inserted in rowid-ascending sequence during a bulk load. Phase 3e.4 uses this to serialize the index to its B-Tree.

Trait Implementations

impl Clone for SecondaryIndex

fn clone(&self) -> SecondaryIndex

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for SecondaryIndex

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

Formats the value using the given formatter.

impl PartialEq for SecondaryIndex

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.