How-to use the Azure CLI in a Bash scripting language
Azure CLI reference commands can execute in several scripting languages. If you're new to Bash and also the Azure CLI, this article a great place to begin your learning journey. Work through this article much like you would a tutorial to learn how to use the Azure CLI in a Bash scripting language with ease.
In this article, you learn how to:
- Query results as JSON dictionaries or arrays
- Format output as JSON, table, or TSV
- Query, filter, and format single and multiple values
- Use if/exists/then and case syntax
- Use for loops
- Use grep, sed, paste, and bc commands
- Populate and use shell and environment variables
If you don't have an Azure subscription, create an Azure free account before you begin.
Starting Bash
Start Bash using Azure Cloud Shell or a local install of the Azure CLI. This article assumes that you're running Bash either using Azure Cloud Shell or running Azure CLI locally in a docker container.
Querying dictionary results
A command that always returns only a single object returns a JSON dictionary. Dictionaries are unordered objects accessed with keys. For this article, we're going to start by querying the Account object using the Account Show command.
az account show
az account show --output json # JSON is the default format
The following JSON dictionary output has some fields omitted for brevity, and identifying information is removed.
bash-5.1# az account show
{
"environmentName": "AzureCloud",
"isDefault": true,
"managedByTenants": [],
"name": "My test subscription",
"state": "Enabled",
"user": {
"name": "user@contoso.com",
"type": "user"
}
}
Formatting the output as YAML
Use the --output yaml
argument (or -o yaml
) to format the output in yaml
format, a plain-text data serialization format. YAML tends to be easier to read than JSON, and
easily maps to that format. Some applications and CLI commands take YAML as configuration input,
instead of JSON.
az account show --output yaml
For more information about formatting the output as yaml, see YAML output format.
Formatting the output as a table
Use the --output table
argument (or -o table
) to format the output as an ASCII table. Nested
objects aren't included in table output, but can still be filtered as part of a query.
az account show --output table
For more information about formatting the output as a table, see Table output format.
Querying and formatting single values and nested values
The following queries demonstrate querying single values, including nested values in a JSON
dictionary output. The final query in this set demonstrates formatting the output using the -o tsv
argument. This argument returns the results as tab- and newline-separated values. This action is
useful for removing quotation marks in the value returned - which is useful to consume the output
into other commands and tools that need to process the text in some form (as is demonstrated later
in this article).
az account show --query name # Querying a single value
az account show --query name -o tsv # Removes quotation marks from the output
az account show --query user.name # Querying a nested value
az account show --query user.name -o tsv # Removes quotation marks from the output
Querying and formatting properties from arrays
The following query demonstrates getting properties in a JSON array. Get subscription properties, displayed as a table of subscriptions.
az account list --query "[].{subscription_id:id, name:name, isDefault:isDefault}" -o table
This query returns results similar to:
Subscription_id Name IsDefault
------------------------------------ ------------------------------------------------- -----------
11111111-3ddc-45ce-8334-c7b28a9e1c3a C & L Azure developer experience content projects False
22222222-8f1c-409b-af1e-8e2e65d9b90a DevCenter - Infrastructure - Dogfood False
33333333-c080-42a7-8973-1aa853ab4df3 Babel False
Querying and formatting multiple values, including nested values
To get more than one property, put expressions in square brackets [ ] (a multiselect list) as a comma-separated list. The following queries demonstrate querying multiple values in a JSON dictionary output, using multiple output formats.
az account show --query [name,id,user.name] # return multiple values
az account show --query [name,id,user.name] -o table # return multiple values as a table
For more information about returning multiple values, see Get multiple values.
Renaming properties in a query
The following queries demonstrate the use of the { } (multiselect hash) operator to get a dictionary instead of an array when querying for multiple values. It also demonstrates renaming properties in the query result.
az account show --query "{SubscriptionName: name, SubscriptionId: id, UserName: user.name}" # Rename the values returned
az account show --query "{SubscriptionName: name, SubscriptionId: id, UserName: user.name}" -o table # Rename the values returned in a table
For more information on renaming properties in a query, see Rename properties in a query.
Querying boolean values
Boolean values are assumed to be true, so the "[?isDefault]"
query syntax for the
az account list
command returns the current default subscription. To get the false values, you
must use an escape character, such as \
.
The following queries demonstrate querying all accounts in a subscription, potentially returning a JSON array if there are multiple subscriptions for a given account, and then querying for which account is the default subscription. It also demonstrates querying for the accounts that aren't the default subscription. These queries build on what you learned previously to filter and format the results. Finally, the final query demonstrates storing the query results in a variable.
az account list
az account list --query "[?isDefault]" # Returns the default subscription
az account list --query "[?isDefault]" -o table # Returns the default subscription as a table
az account list --query "[?isDefault].[name,id]" # Returns the name and id of the default subscription
az account list --query "[?isDefault].[name,id]" -o table # Returns the name and id of the default subscription as a table
az account list --query "[?isDefault].{SubscriptionName: name, SubscriptionId: id}" -o table # Returns the name and id of the default subscription as a table with friendly names
az account list --query "[?isDefault == \`false\`]" # Returns all non-default subscriptions, if any
az account list --query "[?isDefault == \`false\`].name" -o table # Returns all non-default subscriptions, if any, as a table
az account list --query "[?isDefault].id" -o tsv # Returns the subscription id without quotation marks
subscriptionId="$(az account list --query "[?isDefault].id" -o tsv)" # Captures the subscription id as a variable.
echo $subscriptionId # Returns the contents of the variable.
az account list --query "[? contains(name, 'Test')].id" -o tsv # Returns the subscription id of a non-default subscription containing the substring 'Test'
subscriptionId="$(az account list --query "[? contains(name, 'Test')].id" -o tsv) # Captures the subscription id as a variable.
az account set -s $subscriptionId # Sets the current active subscription
- For more information about filtering arrays and querying boolean values, see Filter arrays with boolean expressions.
- For more information about using variables, see How to use variables.
- For more information on working with subscriptions, see Managing subscriptions.
Creating objects using variables and randomization
Setting a random value for use in subsequent commands
Setting and using a random value for use in variables allows you to run scripts multiple times without naming conflicts. Naming conflicts occur because a value must be unique across the service, or because an object you deleted still exists within Azure until the deletion process is complete.
$RANDOM
is a bash function (not a constant) that returns a random signed 16-bit integer (from 0
through 32767). The let
command is a built-in Bash command to evaluate arithmetic expressions.
Using the following command creates a sufficiently unique value for most purposes.
let "randomIdentifier=$RANDOM*$RANDOM"
Working with spaces and quotation marks
Spaces are used for separating commands, options, and arguments. Use quote marks to tell the Bash shell to ignore all special characters, of which a white space is a special character. When the Bash shell sees the first quote mark, it ignores special characters until the closing quote mark. However, sometimes you want the Bash shell to parse certain special characters, such as dollar signs, back quotes, and backslashes. For this scenario, use double quotes.
The following commands use the az group create command to illustrate the use of single and double quote marks. These commands are used to handle spaces and evaluate special characters when working with variables and creating an object.
resourceGroup='msdocs-learn-bash-$randomIdentifier'
echo $resourceGroup # The $ is ignored in the creation of the $resourceGroup variable
resourceGroup="msdocs-learn-bash-$randomIdentifier"
echo $resourceGroup # The $randomIdentifier is evaluated when defining the $resourceGroup variable
location="East US" # The space is ignored when defining the $location variable
echo The value of the location variable is $location # The value of the $location variable is evaluated
echo "The value of the location variable is $location" # The value of the $location variable is evaluated
echo "The value of the location variable is \$location" # The value of the $location variable is not evaluated
echo 'The value of the location variable is $location' # The value of the $location variable is not evaluated
az group create --name $resourceGroup --location $location # Notice that the space in the $location variable is not ignored and the command fails as it treats the value after the space as a new command
az group create --name $resourceGroup --location "$location" # Notice that the space in the $location variable is ignored and the location argument accepts the entire string as the value
In the JSON dictionary output, review the properties of the resource group that was created.
Using If Then Else to determine if variable is null
To evaluate strings, use !=
and to evaluate numbers use -ne
. The following If Then Else statement evaluates whether the $resourceGroup variable has been set. If yes, it returns the value of the variable. If no, it sets the variable.
if [ $resourceGroup != '' ]; then
echo $resourceGroup
else
resourceGroup="msdocs-learn-bash-$randomIdentifier"
fi
Using If Then to create or delete a resource group
The following script creates a new resource group only if one with the specified name doesn't already exist.
if [ $(az group exists --name $resourceGroup) = false ]; then
az group create --name $resourceGroup --location "$location"
else
echo $resourceGroup
fi
The following script deletes an existing new resource group if one with the specified name already
exists. You could use the --no-wait
argument to return control without waiting for the command to
complete. However, for this article, we want to wait for the resource group to be deleted before
continuing. For more information on asynchronous operations, see Tips for using the Azure CLI successfully - Asynchronous operations.
We demonstrate the use of the --no-wait
argument at the end of this article.
if [ $(az group exists --name $resourceGroup) = true ]; then
az group delete --name $resourceGroup -y # --no-wait
else
echo The $resourceGroup resource group does not exist
fi
Using Grep to determine if a resource group exists, and create the resource group if it doesn't
The following command pipes the output of the az group list
command to the grep
command. If the
specified resource group doesn't exist, the command creates the resource group using the previously
defined variables.
az group list --output tsv | grep $resourceGroup -q || az group create --name $resourceGroup --location "$location"
Using CASE statement to determine if a resource group exists, and create the resource group if it doesn't
The following CASE statement creates a new resource group only if one with the specified name doesn't already exist. If one with the specified name exists, the CASE statement echoes that the resource group exists.
var=$(az group list --query "[? contains(name, '$resourceGroup')].name" --output tsv)
case $resourceGroup in
$var)
echo The $resourceGroup resource group already exists.;;
*)
az group create --name $resourceGroup --location "$location";;
esac
Using for loops and querying arrays
In this section of the article, we create a storage account and then use for loops to create blobs and containers. We also demonstrate querying JSON arrays and working with environment variables.
Create storage account
The following command uses the az storage account create command to create a storage account that we use when creating storage containers.
storageAccount="learnbash$randomIdentifier"
az storage account create --name $storageAccount --location "$location" --resource-group $resourceGroup --sku Standard_LRS --encryption-services blob
Get the storage account keys
The following commands use the az storage account keys list command to return storage account key values. We then store a key value in a variable for use when creating storage containers.
az storage account keys list --resource-group $resourceGroup --account-name $storageAccount --query "[].value" -o tsv # returns both storage account key values
az storage account keys list --resource-group $resourceGroup --account-name $storageAccount --query "[0].value" -o tsv # returns a single storage account key value
accountKey=$(az storage account keys list --resource-group $resourceGroup --account-name $storageAccount --query "[0].value" -o tsv)
echo $accountKey
Create storage container
We start by using the az storage container create to create a single storage container and then use the az storage container list to query the name of the created container.
container="learningbash"
az storage container create --account-name $storageAccount --account-key $accountKey --name $container
az storage container list --account-name $storageAccount --account-key $accountKey --query [].name
Upload data to container
The following script creates three sample files using a for loop.
for i in `seq 1 3`; do
echo $randomIdentifier > container_size_sample_file_$i.txt
done
The following script uses the az storage blob upload-batch command to upload the blobs to the storage container.
az storage blob upload-batch \
--pattern "container_size_sample_file_*.txt" \
--source . \
--destination $container \
--account-key $accountKey \
--account-name $storageAccount
The following script uses the az storage blob list command to list the blobs in the container.
az storage blob list \
--container-name $container \
--account-key $accountKey \
--account-name $storageAccount \
--query "[].name"
The following script displays the total bytes in the storage container.
bytes=`az storage blob list \
--container-name $container \
--account-key $accountKey \
--account-name $storageAccount \
--query "[*].[properties.contentLength]" \
--output tsv | paste -s -d+ | bc`
echo "Total bytes in container: $bytes"
echo $bytes
Create many containers using loops
Next, we create multiple containers using a loop demonstrating a couple of ways to write the loop.
for i in `seq 1 4`; do
az storage container create --account-name $storageAccount --account-key $accountKey --name learnbash-$i
done
for value in {5..8}
for (( i=5; i<10; i++));
do
az storage container create --account-name $storageAccount --account-key $accountKey --name learnbash-$i
done
az storage container list --account-name $storageAccount --account-key $accountKey --query [].name
Use EXPORT to define environment variables
In the preceding storage container scripts, we specified the account name and account key with every
command. Instead, you can store your authentication credentials using the corresponding environment
variables: AZURE_STORAGE_ACCOUNT
and AZURE_STORAGE_KEY
. To perform this action, use EXPORT.
export AZURE_STORAGE_ACCOUNT=$storageAccount
export AZURE_STORAGE_KEY=$accountKey
az storage container list # Uses the environment variables to display the list of containers.
The following script creates a metadata string and then uses the az storage container metadata update command to update a container with that string, again using the environment variables.
metadata="key=value pie=delicious" # Define metadata
az storage container metadata update \
--name $container \
--metadata $metadata # Update the metadata
az storage container metadata show \
--name $containerName # Show the metadata
The following command uses the az storage container delete command to delete a single named container and then delete multiple containers in a loop.
az storage container delete \
--name $container
Get list of containers containing a specific prefix and store results into a variable.
containerPrefix="learnbash"
containerList=$(az storage container list \
--query "[].name" \
--prefix $containerPrefix \
--output tsv)
Delete the list of containers in a loop using the --prefix
argument.
for row in $containerList
do
tmpName=$(echo $row | sed -e 's/\r//g')
az storage container delete \
--name $tmpName
done
Error handling
To exit a script immediately if a command returns a nonzero status, run the following command:
set -e
For more information about setting shell options and other help, run the following commands:
help set
help help
Clean up resources
After you complete this article, delete the resource group and all resources within it. Use the
--no-wait
argument.
if [ $(az group exists --name $resourceGroup) = true ]; then
az group delete --name $resourceGroup -y --no-wait
else
echo The $resourceGroup resource group does not exist
fi