Run queries

APPLIES TO: PostgreSQL

Prerequisites

To follow this quickstart, you'll first need to:

  1. Create a cluster in the Azure portal.
  2. Connect to the cluster with psql to run SQL commands.
  3. Create and distribute tables with our example dataset.

Distributed queries

Now it's time for the fun part in our quickstart series--running queries. Let's start with a simple count (*) to verify how much data we loaded in the previous section.

-- count all rows (across shards)

SELECT count(*) FROM github_users;
 count
--------
 264308
(1 row)

Recall that github_users is a distributed table, meaning its data is divided between multiple shards. Azure Cosmos DB for PostgreSQL automatically runs the count on all shards in parallel, and combines the results.

Let's continue looking at a few more query examples:

-- Find all events for a single user.
-- (A common transactional/operational query)

SELECT created_at, event_type, repo->>'name' AS repo_name
  FROM github_events
 WHERE user_id = 3861633;
     created_at      |  event_type  |              repo_name
---------------------+--------------+--------------------------------------
 2016-12-01 06:28:44 | PushEvent    | sczhengyabin/Google-Image-Downloader
 2016-12-01 06:29:27 | CreateEvent  | sczhengyabin/Google-Image-Downloader
 2016-12-01 06:36:47 | ReleaseEvent | sczhengyabin/Google-Image-Downloader
 2016-12-01 06:42:35 | WatchEvent   | sczhengyabin/Google-Image-Downloader
 2016-12-01 07:45:58 | IssuesEvent  | sczhengyabin/Google-Image-Downloader
(5 rows)

More complicated queries

Here's an example of a more complicated query, which retrieves hourly statistics for push events on GitHub. It uses PostgreSQL's JSONB feature to handle semi-structured data.

-- Querying JSONB type. Query is parallelized across nodes.
-- Find the number of commits on the default branch per hour 

SELECT date_trunc('hour', created_at) AS hour,
       sum((payload->>'distinct_size')::int) AS num_commits
FROM   github_events
WHERE  event_type = 'PushEvent' AND
       payload @> '{"ref":"refs/heads/master"}'
GROUP BY hour
ORDER BY hour;
        hour         | num_commits
---------------------+-------------
 2016-12-01 05:00:00 |       13051
 2016-12-01 06:00:00 |       43480
 2016-12-01 07:00:00 |       34254
 2016-12-01 08:00:00 |       29307
(4 rows)

Azure Cosmos DB for PostgreSQL combines the power of SQL and NoSQL datastores with structured and semi-structured data.

In addition to running queries, Azure Cosmos DB for PostgreSQL also applies data definition changes across the shards of a distributed table:

-- DDL commands that are also parallelized

ALTER TABLE github_users ADD COLUMN dummy_column integer;

Next steps

You've successfully created a scalable cluster, created tables, distributed them, loaded data, and run distributed queries.

Now you're ready to learn to build applications with Azure Cosmos DB for PostgreSQL.