In this article, we will look at how we can handle a ‘for loop’ in Terraform. Meta-arguments such as count
and for_each
can be used for different use cases when you want to repeat a set of actions multiple times.
We will look at each one in turn with examples, before looking at how to use the Terraform for
expression with an if
clause. Let’s dive in!
You will learn the following:
- What is a Terraform for loop?
- Types of loops in Terraform
- Terraform count, for_each, and for comparison
- Benefits of using loops in Terraform
- Terraform loops examples
- The for expression with if clause
- The for_each expression with if clause
- Nested Terraform for loops
- Terraform for loop with a map
- How do I iterate over a list in Terraform?
- Terraform for loop best practices
A Terraform for
loop is pretty similar to a loop in any programming language. The key difference, however, is that you can use loops only when building maps or listing comprehension expressions.
The for
expression can be used to transform values and can be used with the if
clause to include or exclude expressions based on a boolean condition.
By default, a resource block configures one object. Using a loop, you can manage several similar objects without writing a separate block for each one. This reduces the amount of code you need to write and makes your scripts cleaner.
Two meta-arguments can be used to do this in Terraform:
count
– This looping construct creates a fixed number of resources based on a count value. (More on Terraform count.)for_each
– This looping construct allows you to create multiple instances of a resource based on a set of input values, such as a list or map. (More on Terraform for_each.)
If the resources you are provisioning are identical or nearly identical, then count
is a safe bet. However, if elements of the resources change between the different instances, then for_each
is the way to go.
You can also use the for
expression to transform a value on each item in a list or map. Use for
when you need to iterate over a set of input values and perform some action on each value.
Below you can find the table comparison of Terraform count
, for_each
and for
.
Construct | Type | Description | Use Case |
Count | Meta-Argument | Based on a count value | Resources you are provisioning are identical |
For_each | Meta-Argument | Based on a set of input values | Resources change between the different instances |
For | Expression | Based on a set of input values | Transform a value |
One of the key benefits of using loops in Terraform is the fact that you can easily build powerful expressions that your resources can leverage. It makes your code reusable and, in the end, minimizes the amount of code you have to write to achieve the desired result. Combining for loops with ifs lets you easily reuse variables in many resources.
In a nutshell, Terraform for
loops help with:
- Code reusability
- Dynamic configurations
- Improved scalability
- Easier resource management
- Conditional creation of resources
- Enhanced error handling
- Support for complex data structures (such as nested lists or maps)
Let’s take a look at some examples.
The example below uses the count
meta-argument to loop through a list of storage account names and create a storage account with the name specified for each.
The name argument uses the count.index
expression to access the current index of the loop (starting from 0) and select the storage account name from the storage_account_names
list using the index. The rest of the arguments are the same for each storage account.
variable "storage_account_names" {
type = list(string)
default = ["jackuksstr001", "jackuksstr002", "jackuksstr003"]
}
resource "azurerm_resource_group" "example" {
name = "storage-rg"
location = "UK South"
}
resource "azurerm_storage_account" "my_storage" {
count = length(var.storage_account_names)
name = var.storage_account_names[count.index]
resource_group_name = azurerm_resource_group.example.name
location = azurerm_resource_group.example.location
account_tier = "Standard"
account_replication_type = "GRS"
}
The example below uses a for_each
loop to iterate through a list of the same storage account names and create a storage account with the name specified for each. The rest of the arguments are the same for each storage account.
The result will be the same as the example using count
above.
variable "storage_account_names" {
type = list(string)
default = ["jackuksstr001", "jackuksstr002", "jackuksstr003"]
}
resource "azurerm_resource_group" "example" {
name = "storage-rg"
location = "UK South"
}
resource "azurerm_storage_account" "my_storage" {
for_each = toset(var.storage_account_names)
name = each.value
resource_group_name = azurerm_resource_group.example.name
location = azurerm_resource_group.example.location
account_tier = "Standard"
account_replication_type = "GRS"
}
The example below builds on the previous one and shows how to output a list of storage account IDs from the given list. The for
expression is used to iterate over the storage_account_names
list and retrieve the ID for each storage account instance with the corresponding name.
variable "storage_account_names" {
type = list(string)
default = ["jackuksstr001", "jackuksstr002", "jackuksstr003"]
}
resource "azurerm_resource_group" "example" {
name = "storage-rg"
location = "UK South"
}
resource "azurerm_storage_account" "my_storage" {
for_each = toset(var.storage_account_names)
name = each.value
resource_group_name = azurerm_resource_group.example.name
location = azurerm_resource_group.example.location
account_tier = "Standard"
account_replication_type = "GRS"
}
output "storage_account_names" {
value = [
for storage in var.storage_account_names:
azurerm_storage_account.my_storage.example[storage].id
]
}
A for
expression can also include an if
clause to filter elements from the source variable, producing a value with fewer elements than the source value, and is commonly used to split lists based on a condition.
The syntax looks like the below:
[for VAR in COLLECTION: IF CONDITION_EXPRESSION: VAR]
VAR
is the name of the variable that represents each item in the collection, COLLECTION
is the collection to be filtered, and CONDITION_EXPRESSION
is the boolean expression that determines whether each item should be included in the filtered collection.
In the example below, we use the for
expression with the if
condition to output a list of storage account names that have the account_replication_type
set to GRS
. This example will output the three storage account names provided in the storage_account_names
variable, as they will all have their account_replication_type
set to GRS
.
variable "storage_account_names" {
type = list(string)
default = ["jackuksstr001", "jackuksstr002", "jackuksstr003"]
}
resource "azurerm_resource_group" "example" {
name = "storage-rg"
location = "UK South"
}
resource "azurerm_storage_account" "my_storage" {
count = length(var.storage_account_names)
name = var.storage_account_names[count.index]
resource_group_name = azurerm_resource_group.example.name
location = azurerm_resource_group.example.location
account_tier = "Standard"
account_replication_type = "GRS"
}
locals {
grs_storage_accounts = [for sa in azurerm_storage_account.my_storage: sa if sa.account_replication_type == "GRS"]
}
output "grs_storage_account_names" {
value = [for sa in local.grs_storage_accounts: sa.name]
}
The if
clause can be used to conditionally include or exclude certain expressions based on a boolean condition.
The syntax for using the if
clause in an expression is as follows:
${condition ? true_value : false_value}
In the example below, we use the if
condition to set the account_replication_type
to GRS
if the environment
variable is set to prod
, if it is not, then the account_replication_type
will be set to LRS
.
Because the default value for the environment
variable is set to prod
in the below example, the three storage accounts created using the for_each
loop will all have their account_replication_type
set to GRS
.
variable "storage_account_names" {
type = list(string)
default = ["jackuksstr001", "jackuksstr002", "jackuksstr003"]
}
variable "environment" {
default = "prod"
}
resource "azurerm_resource_group" "example" {
name = "storage-rg"
location = "UK South"
}
resource "azurerm_storage_account" "my_storage" {
for_each = toset(var.storage_account_names)
name = each.value
resource_group_name = azurerm_resource_group.example.name
location = azurerm_resource_group.example.location
account_tier = "Standard"
account_replication_type = "${var.environment == "prod" ? "GRS" : "LRS"}"
}
Read more about conditional expressions in Terraform.
When you are working with complex data structures, you need to use nested for
loops to easily reuse configurations. DevOps engineers like to use YAML files for most of their configurations, and you can use these files as variables in Terraform, too.
Let’s take a look at an example that contains a nested YAML and how we would go for parsing that input:
instances:
instance1:
ami: ami1
shape: t2-micro
env: dev
instance2:
ami: ami2
shape: t2-micro
env: dev
instance3:
ami: ami1
shape: t2-micro
env: stage
instance4:
ami: ami2
shape: t2-micro
env: stage
In this YAML file, we have declared a couple of instances. Two of them will be used for the dev environment, and the other two will be used for the stage environment.
To read the data from the YAML file, we need to use the file function and to transform the data into a map, we will need to use the YAML decode function:
locals {
data = yamldecode(file("./data.yaml"))
}
Next, if we want to take only the instances that are in the dev environment and build a map with them, we will use a series of expressions:
dev = merge([for data in local.data : { for instance_key, instance_value in data : instance_key => instance_value if instance_value.env == "dev" }]...)
Let’s break it down and forget about the merge function for now:
- Initially, we do a
for
loop to enter the instance map and use a list for that (we don’t need key/value pairs at this point). - Next, as we also want to rebuild the data with key-value pairs, we loop using both the key and the value for each instance.
- We rebuild the data only if the value of the environment is dev.
At this point, the dev data would look like this:
dev = [for data in local.data : { for instance_key, instance_value in data : instance_key => instance_value if instance_value.env == "dev" }]
The value of the data would be:
dev = [
{
"instance1" = {
"ami" = "ami1"
"env" = "dev"
"shape" = "t2-micro"
}
"instance2" = {
"ami" = "ami2"
"env" = "dev"
"shape" = "t2-micro"
}
},
]
This is a list of maps and it would be hard to use in a for_each
loop, so we use the merge function with an ellipsis at the end to transform this into a map:
dev = {
"instance1" = {
"ami" = "ami1"
"env" = "dev"
"shape" = "t2-micro"
}
"instance2" = {
"ami" = "ami2"
"env" = "dev"
"shape" = "t2-micro"
}
}
The entire setup that also shows the data is:
locals {
data = yamldecode(file("./data.yaml"))
dev = merge([for data in local.data : { for instance_key, instance_value in data : instance_key => instance_value if instance_value.env == "dev" }]...)
stage = merge([for data in local.data : { for instance_key, instance_value in data : instance_key => instance_value if instance_value.env == "stage" }]...)
}
output "dev" {
value = local.dev
}
output "stage" {
value = local.stage
}
As you’ve seen in the nested for
loops example, you can use a loop over a map too. You can also iterate through both the keys and the values, and if you use a single iterator, it will iterate through the values. Let’s see it in action:
locals {
pet_map = {
cat = {
color = "orange"
age = "7"
}
dog = {
color = "white"
age = "5"
}
}
pet_colors = [for pet in local.pet_map : pet.color]
}
output "pet_colors" {
value = local.pet_colors
}
Apply complete! Resources: 0 added, 0 changed, 1 destroyed.
Outputs:
pet_colors = [
"orange",
"white",
]
In this example, we’ve just iterated through the values. Now, let’s also iterate through the keys, and create a list of the pets:
locals {
pet_map = {
cat = {
color = "orange"
age = "7"
}
dog = {
color = "white"
age = "5"
}
}
pet_type = [for pet, pet_values in local.pet_map : pet]
}
output "pet_colors" {
value = local.pet_type
}
Apply complete! Resources: 0 added, 0 changed, 0 destroyed.
Outputs:
pet_colors = [
"cat",
"dog",
]
When iterating through lists, by using a single iterator, you will get the values one by one, and if you are using two iterators, the first one will be the index, and the other one will be the value.
Let’s take a look at an example that recreates a list with only the even numbers:
locals {
my_numbers = [1, 3, 5, 2, 8, 12, 33, 10]
even_numbers = [for number in local.my_numbers : number if number % 2 == 0]
}
output "even_numbers" {
value = local.even_numbers
}
Apply complete! Resources: 0 added, 0 changed, 0 destroyed.
Outputs:
even_numbers = [
2,
8,
12,
Now, we will create a list with the index positions for the numbers that are even:
locals {
my_numbers = [1, 3, 5, 2, 8, 12, 33, 10]
even_numbers = [for index, number in local.my_numbers : index if number % 2 == 0]
}
output "even_numbers" {
value = local.even_numbers
}
Apply complete! Resources: 0 added, 0 changed, 0 destroyed.
Outputs:
even_numbers = [
3,
4,
5,
7,
]
There are a couple of things you should have in mind when you are using for
loops in Terraform:
- Keep the number of loops you use to a minimum to make your code easier to understand. This way, the scope is kept narrow, and issues will be easier to debug.
For
loops should be used in locals to simplify the resource configuration. Also, local variables can be easily reused if you need to do so.- Test incrementally – when you are designing a complex loop, go step by step, it will be easier to master how to use loops in any scenario
- Document your loops
- Use descriptive variable names
Even if these are not 100% specific to for
loop, for looping in Terraform in general, try to use for_each
instead of count
all over the place and take advantage of dynamic blocks.
Terraform has two meta-arguments for performing for loops, count
and for_each
. count
should be used when you are provisioning multiple resources that are identical or near identical, and for_each
should be used when resources change between different instances.
We encourage you also to explore how Spacelift makes it easy to work with Terraform. If you need any help managing your Terraform infrastructure, building more complex workflows based on Terraform, and managing AWS credentials per run, instead of using a static pair on your local machine, Spacelift is a fantastic tool for this. It supports Git workflows, policy as code, programmatic configuration, context sharing, drift detection, and many more great features right out of the box. You can check it for free, by creating a trial account.
Note: New versions of Terraform are placed under the BUSL license, but everything created before version 1.5.x stays open-source. OpenTofu is an open-source version of Terraform that expands on Terraform’s existing concepts and offerings. It is a viable alternative to HashiCorp’s Terraform, being forked from Terraform version 1.5.6.
Manage Terraform Better with Spacelift
Build more complex workflows based on Terraform using policy as code, programmatic configuration, context sharing, drift detection, resource visualization and many more.
Terraform Loops and Conditionals Cheatsheet
Grab our ultimate cheat sheet PDF to keep your workflows DRY.