SDK records and queries - Devias

Records are addressed by the API name of their object, such as 'person'. Every method returns a promise.

Create, read, update, delete

const ada = await client.records.create('person', { first_name: 'Ada', last_name: 'Lovelace' });

const same = await client.records.get('person', ada.id);

await client.records.update('person', ada.id, { last_name: 'King' });

await client.records.delete('person', [ada.id]);

delete takes a list of ids, so several records can be removed at once. Two more write methods avoid a separate read:

upsert(object, match, values) updates the record whose unique field matches, or creates one, and tells you which it did.
import(object, body) writes the rows of a CSV, reporting any that are invalid by row.

Ask for records

records.list takes a query, built with q(), a chainable builder.

import { q } from '@devias/sdk';

const page = await client.records.list(
  q('person')
    .where((f) => f.contains('first_name', 'a'))
    .orderBy('first_name', 'asc')
    .limit(10)
    .build(),
);

page.items; // the records
page.nextCursor; // pass it to .cursor() for the next page

Step	What it does
`select(...fields)`	Return only these fields.
`where(build)`	Keep records that match. Combine conditions with `f.and` and `f.or`.
`orderBy(field, direction)`	Sort. Call it again to sort by more.
`limit(n)`, `cursor(c)`	Page through results.
`include(relation)`	Also return the records a relation field points at.
`search(text)`	Match a word across the searchable fields.
`count()`	Also return the total number of matches.

Filters

Filters are built with helpers named after what they check: eq, neq, in, notIn, lt, lte, gt, gte, between, contains, startsWith, isEmpty, isNotEmpty, anyOf and allOf. Combine them with and, or and not.

f.and(f.eq('stage', 'customer'), f.or(f.contains('last_name', 'king'), f.isEmpty('email')))

Page through everything

A page carries a nextCursor while there is more. Keep passing it back until it is null.

Other helpers

lookup finds records by what a person types, for pickers.
aggregate counts and totals records, in groups or overall.

Querying records: the same query as JSON.
Metadata and modules: read the schema.