Quickstart: Analyze with Apache Spark

Article
11/20/2024

In this tutorial, you'll learn the basic steps to load and analyze data with Apache Spark for Azure Synapse.

Prerequisites

Make sure you have placed the sample data in the primary storage account.

Create a serverless Apache Spark pool

In Synapse Studio, on the left-side pane, select Manage > Apache Spark pools.
Select New
For Apache Spark pool name enter Spark1.
For Node size enter Small.
For Number of nodes Set the minimum to 3 and the maximum to 3
Select Review + create > Create. Your Apache Spark pool will be ready in a few seconds.

Understand serverless Apache Spark pools

A serverless Spark pool is a way of indicating how a user wants to work with Spark. When you start using a pool, a Spark session is created if needed. The pool controls how many Spark resources will be used by that session and how long the session will last before it automatically pauses. You pay for spark resources used during that session and not for the pool itself. This way a Spark pool lets you use Apache Spark without managing clusters. This is similar to how a serverless SQL pool works.

Analyze NYC Taxi data with a Spark pool

Note

Make sure you have placed the sample data in the primary storage account.

In Synapse Studio, go to the Develop hub.
Create a new notebook.

Create a new code cell and paste the following code in that cell:

%%pyspark
df = spark.read.load('abfss://users@contosolake.dfs.core.windows.net/NYCTripSmall.parquet', format='parquet')
display(df.limit(10))

Modify the load URI, so it references the sample file in your storage account according to the abfss URI scheme.
In the notebook, in the Attach to menu, choose the Spark1 serverless Spark pool that we created earlier.
Select Run on the cell. Synapse will start a new Spark session to run this cell if needed. If a new Spark session is needed, initially it will take about 2 to 5 minutes to be created. Once a session is created, the execution of the cell will take about 2 seconds.
If you just want to see the schema of the dataframe run a cell with the following code:
```
%%pyspark
df.printSchema()
```

Load the NYC Taxi data into the Spark nyctaxi database

Data is available via the dataframe named df. Load it into a Spark database named nyctaxi.

Add a new code cell to the notebook, and then enter the following code:

%%pyspark
spark.sql("CREATE DATABASE IF NOT EXISTS nyctaxi")
df.write.mode("overwrite").saveAsTable("nyctaxi.trip")

Analyze the NYC Taxi data using Spark and notebooks

Create a new code cell and enter the following code.

%%pyspark
df = spark.sql("SELECT * FROM nyctaxi.trip") 
display(df)

Run the cell to show the NYC Taxi data we loaded into the nyctaxi Spark database.

Create a new code cell and enter the following code. We'll analyze this data and save the results into a table called nyctaxi.passengercountstats.

%%pyspark
df = spark.sql("""
   SELECT passenger_count,
       SUM(trip_distance) as SumTripDistance,
       AVG(trip_distance) as AvgTripDistance
   FROM nyctaxi.trip
   WHERE trip_distance > 0 AND passenger_count > 0
   GROUP BY passenger_count
   ORDER BY passenger_count
""") 
display(df)
df.write.saveAsTable("nyctaxi.passengercountstats")

In the cell results, select Chart to see the data visualized.

Next step

Analyze data with dedicated SQL pool

Share via