Record parsing

After a query executes, the RecordParser coerces raw driver output into typed JavaScript values using the field metadata from the DataModel.

Default parser

DefaultRecordParser handles all categories:

Category	JS type
string, uuid, enum	string
integer	number (or bigint if raw is bigint)
decimal	string (preserves precision)
boolean	boolean
date, timestamp, time	Date
json	pass-through
binary	pass-through
array	pass-through if array, else []

Null and undefined values become null. Fields absent from the row are omitted from the result (partial projection support).

Postgres parser

PostgresRecordParser extends the default with:

• int8 / bigint strings -> JS bigint
• json / jsonb strings -> parsed objects
• Invalid JSON strings kept as-is

MySQL parser

MySQLRecordParser extends the default with:

• bigint strings -> JS bigint
• tinyint(1) numeric 0/1 -> JS boolean
• json strings -> parsed objects
• Invalid JSON strings kept as-is

Output formatters

The SDK includes formatters for serializing parsed records:

import { JsonFormatter, CsvFormatter, RawFormatter } from '@biref/scanner';

// JSON (pretty-printed)
const json = new JsonFormatter({ pretty: true }).serialize(rows);

// CSV with header row
const csv = new CsvFormatter({ includeHeader: true }).serialize(rows);

// Raw (util.inspect)
const raw = new RawFormatter().serialize(rows);