KeyValue Service

The KV service allow you to perform atomic operations on the documents of a collection.
Learn more about more mechanisms like durability or locks in the advanced mechanisms page.

const cluster = connect('couchbase://localhost');

const collection = cluster.bucket('store').scope('library').collection('books');

Retrieving a document

Use Collection.get to fetch a document from the database.

const result = await collection.get('docKey');

If the target bucket have replicas, you may want to fetch the document from a replica instead of the master bucket.
By doing this you will be able to fetch the document in case of node failure, before the rebalance is complete.

WARNING

Since the document may come from a replica, its content may be stale.

const result = await collection.getAnyReplica('docKey');

And if you want to see all the copies of the document :

const resultAllReplicas = await collection.getAllReplicas('docKey');

For each of these operations you can specify some options.

const result = await collection.get('docKey', {
  timeout: 80, // milliseconds
  withExpiry: true, // also fetch when the document expires, if ever
});

Sub-document retrieval

You may not be interested in the entire document. In that case you can use the lookupIn method to only retrieve some parts of the document.
You can get a property, check if a property exists and count the number of elements in an array or the number of keys an object has.

import { LookupInSpec } from '@cbjsdev/cbjs';

const result = await collection.lookupIn('docKey', [
  LookupInSpec.get('title'),
  LookupInSpec.exists('lastModifiedBy'),
  LookupInSpec.count('metadata.tags'),
]);

For this operation, only the timeout option is available.

If you have opt-in for the cluster types on the targeted collection, the paths will be type checked and the result will be typed accordingly.

Chainable sub-doc lookups

Cbjs introduce the ability to chain sub-doc operations. Using this syntax also enables path autocompletion :

import { LookupInSpec } from '@cbjsdev/cbjs';

const lookups = collection
  .lookupIn('docKey')
  .get('title')

if (withTags) {
  lookups.get('metadata.tags');
}

return await lookups;

IMPORTANT

When using the classic lookupIn syntax, the request is executed immediately, regardless of it being awaited or not. A chained lookupIn will only perform the request once awaited await lookups or .then() is called on it.

Error handling

During a lookupIn, the function will not throw if the error is related to a LookupInSpec. Instead, the result will include the error, and the value will be undefined.
In the following example, we assume authors contains a single element.

const { content } = await collection
  .lookupIn('book::001')
  .get('title')
  .get('authors[1]');

const [
  title,
  secondAuthor,
] = content;

if (title.error !== null) {
  // handle the error
}

if (secondAuthor.error !== null) {
  // handle the error
}

Throw on spec error

Each sub-document spec may fail individually. For this reason, the lookup return may be filled with errors. You can see below that bookTitle contains a union type, forcing you to check the error for each sub-doc operation.

const { content: [title, secondAuthor] } = await collection
  .lookupIn('book::001')
  .get('title')
  .get('authors[1]');

if (title.error !== null) {
  // handle the error
}

if (secondAuthor.error !== null) {
  // handle the error
}

If you need all the data in order to proceed, Cbjs provides an additional option when performing the request : throwOnSpecError. Using that option, lookupIn will fail if any spec leads to an error.
This frees you from checking individual spec for error.

const { content: [title, secondAuthor] } = await collection
  .lookupIn('book::001', { throwOnSpecError: true })
  .get('title')
  .get('authors[1]');

Look at the type of title and secondAuthor. No more union.
If either path title or authors[1] don't exist at runtime, it will throw an error, so this is type safe.

Writing a document

There are several way to write a document into Couchbase, depending on the behaviour you want.
You can use Collection.insert, Collection.upsert, Collection.replace or Collection.remove.
Inserting will fail if the document already exists, upserting will succeed regardless.
Replacing or removing a document that does not exist will fail.

await collection.insert(
  'docKey',
  { title: 'Hello' },
  {
    expiry: 45, // the doc will expire in 45 seconds
    timeout: 300,
  }
);

TIP

When upserting, if any expiry time has been set on the document, it will be removed.
If you don't want that, use the preserveExpiry option, the expiry time will then be reset, just like when using Collection.touch.

Sub-document mutation

If you want to modify some parts of the documents, as opposed to the whole document, you can mutate only those parts.

await collection
  .mutateIn('docKey')
  .replace('title', 'New title')
  .replace('lastUpdatedAt', Date.now());

One of the key advantages of sub-document mutation is being able to perform an operation that do not conflict with concurrent ones.
Learn more about optimistic locking.

// The mutation will not overwrite any concurrent changes
await collection.mutateIn('book::001').arrayAddUnique('metadata.tags', 'history');

Chainable sub-doc mutations

Cbjs introduce the ability to chain sub-doc mutations. Using this syntax also enables path autocompletion :

const mutations = collection
  .mutateIn('book::001')
  .replace('title', 'New title');

if (newTag) {
  mutations.arrayAppend('metadata.tags', newTag);
}

return await mutations;

IMPORTANT

When using the classic mutateIn syntax, the request is executed immediately, regardless of it being awaited or not. A chained mutateIn will only perform the request once awaited await mutations or .then() is called on it.

Error handling

The whole mutateIn operation will fail if any of the mutations fail. There is no partial mutation.

Transactions

To perform an ACID transaction with Couchbase, you need to perform the following :

cluster.transactions().run(async ctx => {
  const collection = cluster.bucket('storey').scope('library').collection('books')
  const result = await ctx.get(collection, 'book::001');
});

TIP

Cbjs also exposes ctx.exists, which is missing in the official library.

To offer a more consistent experience, Cbjs gives you the following, alternative API :

cluster.transactions().run(async ctx => {
  const collection = ctx.bucket('storey').scope('library').collection('books')
  const { content } = await collection.get('book::001');
});