Quickstart

Get started with Rockset and go from ingesting data to executing a query using a REST API. Using the Rockset Console, you will learn to:

1. Ingest a public dataset
2. Create a collection
3. Run queries against your collection
4. Create and execute a Query Lambda

#Sign up for Rockset

It’s free to get started with Rockset. And, you can always build for free with under 2 GiB of data.

Sign up for a free 14-day Rockset trial at https://console.rockset.com/create. Part of the signup process is verifying your email address: that is a security measure required for all users who do not authenticate through Github or Google.

The verification process will take you to the Rockset console where you will create an organization. Select your instance type and provide us information on your data source and what you plan to build so we can customize your onboarding experience.

#Create a Collection

A collection in Rockset is the conceptual equivalent of a table in the relational world. Generally, you want a 1:1 mapping from your source tables/topics to Rockset collections.

We will create two collections from public datasets hosted on AWS S3. One dataset is a sample of movies and information including their genre, popularity and revenue. Another dataset is a sample of movie ratings by user. Both datasets are publically available.

Click on Collections in the left hand navigation pane and find the Create a New Collection button. Select sample datasets as the data source.

Give your collection a name and optional description. We’ll use collection names movies and movie_ratings for the remainder of the document. Select the dataset movies and a source preview will automatically generate so you can explore the semi-structured JSON data in a tabular form.

You can apply transformations on incoming data, such as masking sensitive fields, and select a retention policy to automatically drop documents after a period of time.

Click to create the collection. You should now see a new collection in state Created. It can take up to a minute for the collection to become Ready at which point you’ll be able to explore the data and run queries against it.

You will want to repeat creating a sample dataset collection to bring movie_ratings data into Rockset.

Once the collections are Ready, you will enter the collection details view. In the collection details view, you can see the number of documents in the collection and the time of the last update. As new records come in, the count of documents will automatically update and you can run queries against the latest records. You can see all of the available fields, the schema that has been inferred from the collection data, as well as some additional information about the occurrences and distribution of your data. You can also inspect some sample documents in your collection.

#Execute a Query

Now we join and query these two collections using SQL. Click on the Query Editor tab from the left hand sidebar to start writing and running SQL queries. We constructed a query to suggest movies to a user based on their genre preference and the movie rating. Since genre is an array field (as a single movie may fit multiple genres), we’ll need to perform an UNNEST to expand this array and create a record for each (genre, movie) pair. We’ll also join against the movie_ratings table to ensure that no previously seen movies are included in this list.

For the following query, we’ll use the Action genre and user 100 — we’ll generalize these literals in our next step.

SELECT

    m.id,

    m.title

FROM

    commons.movies m,

    UNNEST(m.genres) as genres

WHERE

    genres.name = 'Action'

    AND m.id NOT IN (

        SELECT

            r.movieId

        FROM

            commons.movie_ratings r

        WHERE

            r.userId = '100'

    )

ORDER BY

    m.popularity DESC;

#Parameterize a Query

Use parameters to safely specify literal values in your SQL at runtime. We can add parameters to the SQL query to specify the genre and the userID at runtime.

You can toggle between the query results and the parameters in the query editor. Click to add a parameter.

Let’s use the value Action for genre.

We will use the value 100 for userID.

We can tweak the SQL statement to incorporate the parameters we just created. Here’s the new SQL statement with the parameters:

SELECT

    m.id,

    m.title

FROM

    commons.movies m,

    UNNEST(m.genres) as genres

WHERE

    genres.name = :genre

    AND m.id NOT IN (

        SELECT

            r.movieId

        FROM

            commons.movie_ratings r

        WHERE

            r.userId = :userId

    )

ORDER BY

    m.popularity DESC;

#Create a Query Lambda

Let’s use the query with parameters to create a Query Lambda. Click Create Query Lambda. Query Lambdas are named, parameterized SQL queries stored in Rockset that can be executed from a dedicated REST endpoint. Use Query Lambdas to build applications backed by Rockset as opposed to querying with raw SQL directly from application code.

We will use the default parameter values of genre and userId at runtime.

Create an API key and copy the code snippet into the application code.

Open the terminal or your script/application and copy the request to execute the query.

#Invite Other Users

Invite members of your team to Rockset through the console. To invite new members, click on the left hand navigation Users. You can determine if the new users should be administrators, members or have read-only access to Rockset.

#Next Steps

Join us on the Slack community and share what you are looking to build with Rockset. We’re hanging out and ready to answer your questions.

Also check out some of the pages below to continue exploring Rockset:

• Setup an integration to your data source
• Create complex queries specific to your application
• Explore SDKs