A newer version of Hazelcast Platform is available.

View latest

Get Started with SQL Over Files

In this tutorial, you learn the basics of querying files in SQL by creating and querying a CSV file.

Before You Begin

To complete this tutorial, you need the following:

Prerequisites	Useful resources
A Hazelcast cluster in client/server mode and an instance of Management Center running on your local network.	Start a Local Cluster.
A connection to the SQL shell	SQL

Prerequisites

Useful resources

A Hazelcast cluster in client/server mode and an instance of Management Center running on your local network.

Start a Local Cluster.

A connection to the SQL shell

SQL

Step 1. Create a Data Source to Query

SQL can query data in maps, Kafka topics and the local cluster’s file system.

In this step, you create a CSV file that you can use to query.

Create a file named likes.csv.
Add the following data to your file.
```
msg_id,likes,dislikes
1,20,13
2,108,25
3,122,73
4,9,88
5,51,42
```
This example file contains a record of the number of likes each message has.

Step 2. Create a Mapping to the File

To allow Hazelcast to find and recognize the data in your CSV file, you need to create a mapping to it.

In the SQL shell, use the CREATE MAPPING statement to configure the file connector and give Hazelcast access to the data in the likes.csv file.

CREATE MAPPING csv_likes (msg_id INT, likes INT, dislikes INT)
TYPE File
OPTIONS ('format'='csv',
    'path'='/absolute/path/to/current/directory', 'glob'='likes.csv');

Make sure you replace the path option with the absolute path to your CSV file.

Step 3. Run Ad-Hoc Queries

Ad-hoc queries allow you to retrieve a small subset of data. Usually these queries are simple and you can have many of them running concurrently in a Hazelcast cluster.

Use a SELECT statement to query all the data in the likes.csv file.

SELECT * FROM csv_likes;

You should see the following:

+------------+------------+------------+
|      msg_id|       likes|    dislikes|
+------------+------------+------------+
|           1|          20|          13|
|           4|           9|          88|
|           3|         122|          73|
|           2|         108|          25|
|           5|          51|          42|
+------------+------------+------------+

Query only the msg_id and likes columns, by adding them as a comma-separated list after the SELECT statement.

SELECT msg_id, likes FROM csv_likes;

+------------+------------+
|      msg_id|       likes|
+------------+------------+
|           1|          20|
|           3|         122|
|           4|           9|
|           2|         108|
|           5|          51|
+------------+------------+

Use a filter to display only the message numbers with more than 50 likes.

SELECT msg_id FROM csv_likes WHERE likes > 50;

+------------+
|      msg_id|
+------------+
|           2|
|           5|
|           3|
+------------+

Give the msg_id column an alias for the query results.

This clause does not rename the column in the table.

SELECT msg_id AS message_number, likes, dislikes
FROM csv_likes
WHERE likes > 20;

+--------------+------------+------------+
|message_number|       likes|    dislikes|
+--------------+------------+------------+
|             2|         108|          25|
|             3|         122|          73|
|             5|          51|          42|
+--------------+------------+------------+

To filter rows on more than one condition, you can join conditions with the AND, OR, and NOT operators.

SELECT *
FROM csv_likes
WHERE likes > 100 OR dislikes < 30;

+------------+------------+------------+
|      msg_id|       likes|    dislikes|
+------------+------------+------------+
|           1|          20|          13|
|           2|         108|          25|
|           3|         122|          73|
+------------+------------+------------+

SELECT *
FROM csv_likes
WHERE likes > 100 AND dislikes < 30;

+------------+------------+------------+
|      msg_id|       likes|    dislikes|
+------------+------------+------------+
|           2|         108|          25|
+------------+------------+------------+

If you need more control over how your data is being transformed and aggregated, you may want to build a pipeline with the Jet API.

Step 4. Run Federated Queries

Federated queries are those that join tables from different datasets.

Normally, SQL queries are executed on one particular database or dataset. However, with Hazelcast, you can pull information from different sources and present a more complete picture of the data.

Configure the map connector to create a new table called dislikes.
```
CREATE MAPPING names
TYPE IMap OPTIONS ('keyFormat'='int', 'valueFormat'='varchar');
```
This table is mapped to a distributed map in Hazelcast where the key is an integer and the value is a string.

Use the SINK INTO statement to add some entries to the map.

SINK INTO names VALUES
(1, 'Greg'),
(2, 'Jerry'),
(3, 'Mary'),
(4, 'Jerry'),
(5, 'Joe');

Use the JOIN clause to merge results from the messages and names tables so you can see who has the most likes and dislikes.

SELECT names.this, csv_likes.likes, csv_likes.dislikes
FROM csv_likes
JOIN names
ON csv_likes.msg_id = names.__key;

+--------------------+------------+------------+
|this                |       likes|    dislikes|
+--------------------+------------+------------+
|Jerry               |         108|          25|
|Greg                |          20|          13|
|Jerry               |           9|          88|
|Joe                 |          51|          42|
|Mary                |         122|          73|
+--------------------+------------+------------+

Use the ORDER BY clause to order the results by name and use the LIMIT clause to limit them so that only the first two are displayed. Change the header of the names column to name.

SELECT names.this AS name, csv_likes.likes, csv_likes.dislikes
FROM csv_likes
JOIN names
ON csv_likes.msg_id = names.__key
ORDER BY names.this
LIMIT 2;

+--------------------+------------+------------+
|name                |       likes|    dislikes|
+--------------------+------------+------------+
|Greg                |          20|          13|
|Jerry               |           9|          88|
+--------------------+------------+------------+

Use the SUM() function to aggregate the total number of likes for each person and group the results by name.

SELECT names.this AS name, sum(csv_likes.likes) AS total_likes
FROM csv_likes
JOIN names
ON csv_likes.msg_id = names.__key
GROUP BY name;

You should see the following:

+--------------------+--------------------+
|name                |         total_likes|
+--------------------+--------------------+
|Greg                |                  20|
|Mary                |                 122|
|Joe                 |                  51|
|Jerry               |                 117|
+--------------------+--------------------+

The results do not include a row for each Jerry because the GROUP BY statement groups the results by name.

Filter for the names that have more than 100 likes combined, using the HAVING clause. This clause is equivalent to the WHERE clause but for aggregate results.

SELECT names.this AS most_liked
FROM csv_likes
JOIN names
ON csv_likes.msg_id = names.__key
GROUP BY names.this HAVING SUM(likes) > 100;

+--------------------+
|most_liked          |
+--------------------+
|Jerry               |
|Mary                |
+--------------------+

For a list of available aggregations, see Expressions, Functions and Operators.

Step 5. Ingest Query Results into a Hazelcast Map

To save your query results as a view, you can cache them in Hazelcast by ingesting them into a map.

Configure the map connector to create a new table called likes_and_dislikes.
```
CREATE MAPPING likes_and_dislikes (
__key INT,
name VARCHAR,
likes INT,
dislikes INT
) TYPE IMap OPTIONS ('keyFormat'='int', 'valueFormat'='json-flat');
```
This table is mapped to a distributed map in Hazelcast where the key is an integer and the value is an object that’s serialized to JSON.

Run the JOIN query to merge results from the CSV file and the dislikes map and insert them into the likes_and_dislikes map.

INSERT INTO likes_and_dislikes SELECT csv_likes.msg_id, names.this, csv_likes.likes, csv_likes.dislikes
FROM csv_likes
JOIN names
ON csv_likes.msg_id = names.__key;

Make sure that the query results were added to the map.

SELECT * FROM likes_and_dislikes
ORDER BY __key;

+------------+--------------------+------------+------------+
|       __key|name                |       likes|    dislikes|
+------------+--------------------+------------+------------+
|           1|Greg                |          20|          13|
|           2|Jerry               |         108|          25|
|           3|Mary                |         122|          73|
|           4|Jerry               |           9|          88|
|           5|Joe                 |          51|          42|
+------------+--------------------+------------+------------+