surrealdb-core 3.2.1

A scalable, distributed, collaborative, document-graph database, for the realtime web
Documentation
//! Document ID Mapping Key (`Id`) for Full-Text Index
//!
//! The `Id` key stores the mapping between SurrealDB record IDs (`Thing`) and
//! internal numeric document IDs (`DocId`) used by the full-text search engine.
//!
//! ## Key Structure
//! ```no_compile
//! /*{namespace}*{database}*{table}+{index}!id{record_id}
//! ```
//!
//! ## Purpose
//! - **ID Translation**: Converts between user-facing record IDs and internal numeric document IDs
//! - **Bidirectional Mapping**: Works with `Bi` keys to provide reverse lookups
//! - **Index Efficiency**: Numeric document IDs are more efficient for internal search operations
//!
//! ## Usage in Full-Text Search
//! The `Id` key is essential for the full-text search pipeline:
//! 1. **Indexing Phase**: Record IDs are converted to document IDs using `Id` keys
//! 2. **Search Phase**: Results use document IDs internally for efficiency
//! 3. **Result Retrieval**: Document IDs are converted back to record IDs for user presentation
//!
//! ## Category
//! - **Category**: `IndexInvertedDocIds`
//! - **Domain**: Full-text search document ID mapping
//!
//! ## Integration with Document ID Lifecycle
//! 1. **ID Resolution**: When a document is indexed, its record ID is mapped to a numeric document
//!    ID
//! 2. **Storage**: The `Id` key stores: `record_id → doc_id`
//! 3. **Allocation**: If no mapping exists, a new document ID is allocated from the sequence (using
//!    `Ib` keys)
//! 4. **Reverse Mapping**: A complementary `Bi` key stores: `doc_id → record_id`
//!
//! ## Performance Characteristics
//! - **Space Efficient**: Numeric document IDs are smaller than full record IDs
//! - **Cache Friendly**: Sequential numeric IDs improve cache locality
//! - **Concurrent Safe**: Works with distributed sequence mechanism to prevent ID conflicts
//! - **Scalable**: Efficient lookups scale with the number of indexed documents
use std::borrow::Cow;
use std::fmt::Debug;

use storekey::{BorrowDecode, Encode};

use crate::catalog::{DatabaseId, IndexId, NamespaceId};
use crate::idx::seqdocids::DocId;
use crate::key::category::{Categorise, Category};
use crate::val::{IndexFormat, RecordIdKey, TableName};

#[derive(Debug, Clone, PartialEq, Encode, BorrowDecode)]
#[storekey(format = "IndexFormat")]
pub(crate) struct Id<'a> {
	__: u8,
	_a: u8,
	pub ns: NamespaceId,
	_b: u8,
	pub db: DatabaseId,
	_c: u8,
	pub tb: Cow<'a, TableName>,
	_d: u8,
	pub ix: IndexId,
	_e: u8,
	_f: u8,
	_g: u8,
	pub id: RecordIdKey,
}

impl crate::kvs::KVKey for Id<'_> {
	type ValueType = DocId;

	fn encode_key(&self) -> anyhow::Result<Vec<u8>> {
		Ok(storekey::encode_vec_format::<IndexFormat, _>(self)
			.map_err(|_| crate::err::Error::Unencodable)?)
	}

	fn value_context(&self) {}
}

impl Categorise for Id<'_> {
	fn categorise(&self) -> Category {
		Category::IndexInvertedDocIds
	}
}

impl<'a> Id<'a> {
	#[cfg_attr(target_family = "wasm", allow(dead_code))]
	pub fn new(
		ns: NamespaceId,
		db: DatabaseId,
		tb: &'a TableName,
		ix: IndexId,
		id: RecordIdKey,
	) -> Self {
		Self {
			__: b'/',
			_a: b'*',
			ns,
			_b: b'*',
			db,
			_c: b'*',
			tb: Cow::Borrowed(tb),
			_d: b'+',
			ix,
			_e: b'!',
			_f: b'i',
			_g: b'd',
			id,
		}
	}
}

#[cfg(test)]
mod tests {
	use super::*;
	use crate::kvs::KVKey;

	#[test]
	fn key() {
		let tb = TableName::from("testtb");
		let val = Id::new(
			NamespaceId(1),
			DatabaseId(2),
			&tb,
			IndexId(3),
			RecordIdKey::from("id".to_owned()),
		);
		let enc = Id::encode_key(&val).unwrap();
		assert_eq!(
			enc,
			b"/*\x00\x00\x00\x01*\x00\x00\x00\x02*testtb\0+\0\0\0\x03!id\x03id\0",
			"{}",
			String::from_utf8_lossy(&enc)
		);
	}
}