Indexes & explain

Indexes are created per collection field and are persisted on disk.

Create an index

createIndex(collection: string, field: string, options?: { unique?: boolean }): Promise<void>

• Builds the index from existing documents.
• After creation, writes keep the index updated.
• unique: true enforces uniqueness for that field.

TypeScript

import { LioranManager } from "@liorandb/core";

const manager = new LioranManager({ rootPath: "./data" });
const db = await manager.db("app");

db.collection("users"); // ensure collection exists
await db.createIndex("users", "email", { unique: true });

const users = db.collection("users");
console.log(await users.findOne({ email: { $eq: "a@b.com" } }));

await manager.close();

JavaScript

import { LioranManager } from "@liorandb/core";

const manager = new LioranManager({ rootPath: "./data" });
const db = await manager.db("app");

db.collection("users");
await db.createIndex("users", "email", { unique: true });

const users = db.collection("users");
console.log(await users.findOne({ email: { $eq: "a@b.com" } }));

await manager.close();

How indexes are used

• Query execution tries to pick one best index from your query.
• If no index can be selected, a full scan is used.
• Even with an index, documents are still re-checked against your full query to ensure correctness.

Supported indexed shapes include:

• { field: value }
• { field: { $eq: value } }
• { field: { $in: [...] } }
• { field: { $gt/$gte/$lt/$lte: ... } } (range queries)

explain()

You can ask for query plan stats:

db.explain("users", { age: { $gte: 18 } })

The result includes:

• indexUsed: which field index was used (if any)
• scannedDocuments: how many documents were read
• candidateDocuments: how many IDs the index/full-scan produced
• executionTimeMs: elapsed time in milliseconds