Databricks Lab 1

Analyzing Data with Spark in Azure
Databricks
Lab 1 - Getting Started with Spark
Overview
In this lab, you will provision a Databricks workspace and a Spark cluster. You will then use the Spark
cluster to explore data interactively.
What You’ll Need

To complete the labs, you will need the following:
• A web browser
• A Microsoft account
• A Microsoft Azure subscription
• A Windows, Linux, or Mac OS X computer
• Azure Storage Explorer
• The lab files for this course
Note: To set up the required environment for the lab, follow the instructions in the Setup document for
this course. Specifically, you must have signed up for an Azure subscription.
Provisioning Azure Resources

Note: If you already have an Azure Databricks Spark cluster and an Azure blob storage account, you can
skip this section.
Provision a Databricks Workspace

1. In a web browser, navigate to http://portal.azure.com, and if prompted, sign in using the
Microsoft account that is associated with your Azure subscription.
2. In the Microsoft Azure portal, click  Create a resource. Then in the Analytics section select
Azure Databricks and create a new Azure Databricks workspace with the following settings:
• Workspace name: Enter a unique name (and make a note of it!)
• Subscription: Select your Azure subscription
• Resource Group: Create a new resource group with a unique name (and make a note of
it!)
• Location: Choose any available data center location.
• Pricing Tier: Standard
3. In the Azure portal, view Notifications to verify that deployment has started. Then wait for the
workspace to be deployed (this can take few minutes)
Provision a Storage Account
1. In the Azure portal tab in your browser, and click  Create a resource.
2. In the Storage category, click Storage account.
3. Create a new storage account with the following settings:
• Name: Specify a unique name (and make a note of it)
• Deployment model: Resource manager
• Account kind: Storage (general purpose v1)
• Location: Choose the same location as your Databricks workspace
• Replication: Locally-redundant storage (LRD)
• Performance: Standard
• Secure transfer required: Disabled
• Subscription: Choose your Azure subscription
• Resource group: Choose the existing resource group for your Databricks workspace
• Virtual networks: Disabled
4. Wait for the resource to be deployed. Then view the newly deployed storage account.
5. In the blade for your storage account, click Blobs.
6. In the Browse blobs blade, click  Container, and create a new container with the following
settings:
• Name: spark
• Public access level: Private (no anonymous access)
7. In the Settings section of the blade for your blob store, click Access keys and note the Storage
account name and key1 values on this blade – you will need these in the next procedure.
Create a Spark Cluster

1. In the Azure portal, browse to the Databricks workspace you created earlier, and click Launch
Workspace to open it in a new browser tab.
2. In the Azure Databricks workspace home page, under New, click Cluster.
3. In the Create Cluster page, create a new cluster with the following settings:
• Cluster Mode: Standard
• Cluster Name: Enter a unique cluster name (and make a note of it)
• Databricks Runtime Version: Choose the latest available version
• Python Version: 3
• Driver Type: Same as worker
• Worker Type: Leave the default type
• Min Workers: 1
• Max Workers: 2
• Auto Termination: Terminate after 60 minutes.
• Spark Config: Add two key-value pairs for your storage account and key like this:
fs.azure.account.key.your_storage_account.blob.core.windows.net your_key1_value
spark.hadoop.fs.azure.account.key.your_storage_account.blob.core.windows.net your_key1_value
Note: The first setting enables code using the newer Dataframe-based API to access your
storage account. The second setting is used by the older RDD-based API.
4. Wait for the cluster to be created.
Exploring Data Interactively with Spark RDDs
Now that you have provisioned a Spark cluster, you can use it to analyze data. In this exercise, you will
use Spark Resilient Distributed Datasets (RDDs) to load and explore data. The RDD-based API is an
original component of Spark, and has largely been superseded by a newer Dataframe-based API;
however, there are many production systems (and code examples on the Web) that use RDDs, so it’s
worth starting your exploration of Spark there.
Upload Source Data to Azure Storage

In this exercise, you will use the Spark RDD API to explore unstructured data. Before you can do this, you
must store the data files you want to explore in a blob storage container where it can be accessed by
your cluster. The instructions here assume you will use Azure Storage Explorer to do this, but you can
use any Azure Storage tool you prefer.
1. In the folder where you extracted the lab files for this course on your local computer, in the data
folder, verify that the KennedyInaugural.txt file exist. This file contains the data you will explore
in this exercise.
2. Start Azure Storage Explorer, and if you are not already signed in, sign into your Azure
subscription.
3. Expand your storage account and the Blob Containers folder, and then double-click the spark
blob container you created previously.
4. In the Upload drop-down list, click Upload Files. Then upload KennedyInaugural.txt as a block
blob to a new folder named data in root of the spark container.
Create a Notebook
Most interactive data analysis in Databricks is conducted using notebooks. These browser-based
interactive documents enable you to combine notes in Markdown format with code that you can run
right in the notebook – with no requirement to install a local code editing environment. In this exercise,
you can choose to write your code using Python or Scala.
1. In the Databricks workspace, click Workspace. Then click Users, click your user name, and in the
drop-down menu for your username point click Create and Notebook as shown here:
2. Create a new notebook with the following settings:
• Name: RDDs
• Language: Choose Python or Scala as preferred.
• Cluster: Your cluster
3. In the new notebook, in the first cell, enter the following code to enter some Markdown text:
%md
# Kennedy Inauguration
This notebook contains code to analyze President Kennedy’s inauguration speech.
4. Click anywhere in the notebook outside of the first cell to see the formatted markdown, which
should look like this:
Kennedy Inauguration
This notebook contains code to analyze President Kennedy’s inauguration speech.
5. Hold the mouse pointer under the center of the bottom edge of the cell until a (+) symbol is
displayed; then click this to insert a new cell.
6. In the new cell, type the following code, replacing <account> with the fully qualified name of
your Azure Storage account (account_name.blob.core.windows.net):
Python
txt = sc.textFile("wasbs://spark@<account>/data/KennedyInaugural.txt")
txt.count()
Scala
val txt = sc.textFile("wasbs://spark@<account>/data/KennedyInaugural.txt")
txt.count()
In this code, the variable sc is the Spark context for your cluster; which is created automatically
within the notebook.
7. With the code cell selected, at the top left of the cell, click the  button and then click  Run
Cell to run the cell. After a few seconds, the code will run and display the number of lines of text
in the text file as Out[1]
8. Add a new cell and enter the following command to view the first line in the text file.
Python
txt.first()
Scala
txt.first()
9. Run the new cell and note that the first line of the speech is displayed as Out[2].
10. Add a new cell and enter the following command to create a new RDD named filtTxt that filters
the txt RDD so that only lines containing the word “freedom” are included, and counts the
filtered lines.
Python
filtTxt = txt.filter(lambda line: "freedom" in line)
filtTxt.count()
Scala
val filtTxt = txt.filter(line => line.contains("freedom"))
filtTxt.count()
11. Run the new cell and note that the number of lines containing “freedom” is returned as Out[3].
12. Add a new cell and enter the following command to display the contents of the filtTxt RDD.
Python
filtTxt.collect()
Scala
filtTxt.collect()
13. Run the new cell and note that the lines containing “freedom” are returned as Out[4].
14. Add a new cell and enter the following command to split the full speech into words, count the
number of times each word occurs, and display the counted words in descending order of
frequency.
Python
words = txt.flatMap(lambda txt: txt.split(" "))
counts = words.map(lambda word: (word, 1)).reduceByKey(lambda a, b: a + b)
counts.sortBy(lambda a: a[1], False).collect()
Scala
val words = txt.flatMap(line => line.split(" "))
val counts = words.map(word => (word, 1)).reduceByKey((a, b) => a + b)
counts.sortBy(_._2,false).collect().foreach(println)
15. Run the new cell and review the output, which shows the frequency of each word in the speech
in descending order.
Exploring Data Interactively with Dataframes

Spark 2.0 and later provides a schematized object for manipulating and querying data – the DataFrame.
This provides a much more intuitive, and better performing, API for working with structured data. In
addition to the native Dataframe API, Spark SQL enables you to use SQL semantics to create and query
tables based on Dataframes.
Upload Source Data to Azure Storage

In this lab, you will explore both structured data relating to road traffic accidents. Before you can do
this, you must upload the data files containing the data to your blob storage container where it can be
accessed by your cluster. The instructions here assume you will use Azure Storage Explorer to do this,
but you can use any Azure Storage tool you prefer.
1. In the folder where you extracted the lab files for this course on your local computer, in the data
folder, verify that the Accidents.csv and Vehicles.csv files exist. These files contain the data you
will explore in this exercise.
2. Start Azure Storage Explorer, and if you are not already signed in, sign into your Azure
subscription.
3. Expand your storage account and the Blob Containers folder, and then double-click the spark
blob container you created previously in this lab.
4. In the Upload drop-down list, click Upload Files. Then upload Accidents.csv and Vehicles.csv as
block blobs to the data folder in root of the spark container that you created previously in this
lab.
Work with Dataframes
In this procedure, you will use your choice of Python or Scala to query the road traffic accident data
in the comma-delimited text files you have uploaded. Notebooks containing the necessary steps to
explore the data have been provided.
1. In the Databricks workspace, click Workspace. Then click Users, click your user name, and in the
drop-down menu for your username click Import as shown here:
2. Browse to the folder where you extracted the lab files. Then select either Dataframes.ipynb or
Dataframes.scala, depending on your preferred choice of language (Python or Scala), and
upload it.
3. Open the notebook you uploaded and in the Detached drop-down menu, attach the notebook
to your Spark cluster as shown here:
4. Read the notes and run the code cells to explore the data.
Clean Up
Note: If you intend to proceed straight to the next lab, skip this section. Otherwise, follow the steps
below to delete your Azure resources and avoid being charged for them when you are not using them.
Delete the Resource Group

1. Close the browser tab containing the databricks workspace if it is open.
2. In the Azure portal, view your Resource groups and select the resource group you created for
your databricks workspace. This resource group contains your databricks workspace and your
storage account.
3. In the blade for your resource group, click Delete. When prompted to confirm the deletion,
enter the resource group name and click Delete.
4. Wait for a notification that your resource group has been deleted.
5. After a few minutes, a second resource group containing the resources for your cluster will
automatically be deleted.
6. Close the browser.

Databricks Lab 1

Uploaded by

Copyright:

Available Formats

Databricks Lab 1

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Databricks Lab 1

Uploaded by

Copyright:

Available Formats

What are the steps to provision an Azure Databricks workspace and Spark cluster?

What are the different ways to explore data using Spark?

Analyzing Data with Spark in Azure

What You’ll Need

Provisioning Azure Resources

Provision a Databricks Workspace

Create a Spark Cluster

Upload Source Data to Azure Storage

Exploring Data Interactively with Dataframes

Upload Source Data to Azure Storage

Delete the Resource Group

You might also like