Module table

Structs§

Column: The schema for each SQL column in every table.
FtsIndexEntry: One FTS index attached to a table (Phase 8b). The inverted index itself is a PostingList; metadata (name, column, dirty flag) lives here. Mirrors HnswIndexEntry field-for-field so the rebuild-on-save and DELETE/UPDATE invalidation paths can use one pattern across both index families.
HnswIndexEntry: One HNSW index attached to a table. The distance metric is fixed at CREATE INDEX time via USING hnsw (col) WITH (metric = '<m>') (l2 / cosine / dot); omitting the WITH clause defaults to L2, matching the pre-SQLR-28 behaviour for round-tripping older sqlrite_master rows that didn’t carry a metric.
Table: The schema for each SQL Table is represented in memory by following structure.

DataType: SQLRite data types Mapped after SQLite Data Type Storage Classes and SQLite Affinity Type (Datatypes In SQLite Version 3)[https://www.sqlite.org/datatype3.html]
Row: The schema for each SQL row in every table is represented in memory by following structure
Value: Runtime value produced by query execution. Separate from the on-disk Row enum so the executor can carry typed values (including NULL) across operators.

parse_vector_literal: Parse a bracket-array literal like "[0.1, 0.2, 0.3]" (or "[1, 2, 3]") into a Vec<f32>. The parser/insert pipeline stores vector literals as strings in InsertQuery::rows (a Vec<Vec<String>>); this helper is the inverse — turn the string back into a typed vector at the boundary where we actually need element-typed data.