terraform/website/source/docs/configuration/interpolation.html.md

337 lines
15 KiB
Markdown
Raw Normal View History

2014-07-28 19:43:00 +02:00
---
layout: "docs"
page_title: "Interpolation Syntax"
sidebar_current: "docs-config-interpolation"
2014-10-22 05:21:56 +02:00
description: |-
Embedded within strings in Terraform, whether you're using the Terraform syntax or JSON syntax, you can interpolate other values into strings. These interpolations are wrapped in `${}`, such as `${var.foo}`.
2014-07-28 19:43:00 +02:00
---
# Interpolation Syntax
Embedded within strings in Terraform, whether you're using the
Terraform syntax or JSON syntax, you can interpolate other values
into strings. These interpolations are wrapped in `${}`, such as
`${var.foo}`.
The interpolation syntax is powerful and allows you to reference
variables, attributes of resources, call functions, etc.
You can also perform simple math in interpolations, allowing
you to write expressions such as `${count.index + 1}`.
You can escape interpolation with double dollar signs: `$${foo}`
will be rendered as a literal `${foo}`.
2014-10-03 07:11:53 +02:00
## Available Variables
**To reference user variables**, use the `var.` prefix followed by the
2014-07-28 19:43:00 +02:00
variable name. For example, `${var.foo}` will interpolate the
`foo` variable value. If the variable is a mapping, then you
can reference static keys in the map with the syntax
`var.MAP.KEY`. For example, `${var.amis.us-east-1}` would
get the value of the `us-east-1` key within the `amis` variable
that is a mapping.
2015-02-24 00:04:18 +01:00
**To reference attributes of your own resource**, the syntax is
`self.ATTRIBUTE`. For example `${self.private_ip_address}` will
interpolate that resource's private IP address. Note that this is
only allowed/valid within provisioners.
2014-10-03 07:11:53 +02:00
**To reference attributes of other resources**, the syntax is
2014-07-28 19:43:00 +02:00
`TYPE.NAME.ATTRIBUTE`. For example, `${aws_instance.web.id}`
will interpolate the ID attribute from the "aws\_instance"
resource named "web". If the resource has a `count` attribute set,
you can access individual attributes with a zero-based index, such
as `${aws_instance.web.0.id}`. You can also use the splat syntax
to get a list of all the attributes: `${aws_instance.web.*.id}`.
This is documented in more detail in the
[resource configuration page](/docs/configuration/resources.html).
2014-07-28 19:43:00 +02:00
2014-10-03 07:11:53 +02:00
**To reference outputs from a module**, the syntax is
`MODULE.NAME.OUTPUT`. For example `${module.foo.bar}` will
interpolate the "bar" output from the "foo"
[module](/docs/modules/index.html).
**To reference count information**, the syntax is `count.FIELD`.
For example, `${count.index}` will interpolate the current index
in a multi-count resource. For more information on count, see the
resource configuration page.
2014-07-28 19:43:00 +02:00
<a id="path-variables"></a>
2014-10-08 05:15:08 +02:00
**To reference path information**, the syntax is `path.TYPE`.
TYPE can be `cwd`, `module`, or `root`. `cwd` will interpolate the
cwd. `module` will interpolate the path to the current module. `root`
will interpolate the path of the root module. In general, you probably
want the `path.module` variable.
2014-07-28 19:43:00 +02:00
## Built-in Functions
2014-10-03 07:11:53 +02:00
Terraform ships with built-in functions. Functions are called with
the syntax `name(arg, arg2, ...)`. For example,
to read a file: `${file("path.txt")}`. The built-in functions
are documented below.
2014-07-28 19:43:00 +02:00
The supported built-in functions are:
* `base64decode(string)` - Given a base64-encoded string, decodes it and
returns the original string.
* `base64encode(string)` - Returns a base64-encoded representation of the
given string.
* `base64sha256(string)` - Returns a base64-encoded representation of raw
SHA-256 sum of the given string.
**This is not equivalent** of `base64encode(sha256(string))`
since `sha256()` returns hexadecimal representation.
* `cidrhost(iprange, hostnum)` - Takes an IP address range in CIDR notation
and creates an IP address with the given host number. For example,
``cidrhost("10.0.0.0/8", 2)`` returns ``10.0.0.2``.
* `cidrnetmask(iprange)` - Takes an IP address range in CIDR notation
and returns the address-formatted subnet mask format that some
systems expect for IPv4 interfaces. For example,
``cidrmask("10.0.0.0/8")`` returns ``255.0.0.0``. Not applicable
to IPv6 networks since CIDR notation is the only valid notation for
IPv6.
* `cidrsubnet(iprange, newbits, netnum)` - Takes an IP address range in
CIDR notation (like ``10.0.0.0/8``) and extends its prefix to include an
additional subnet number. For example,
``cidrsubnet("10.0.0.0/8", 8, 2)`` returns ``10.2.0.0/16``;
``cidrsubnet("2607:f298:6051:516c::/64", 8, 2)`` returns
``2607:f298:6051:516c:200::/72``.
2016-01-06 22:19:54 +01:00
2015-11-08 07:34:56 +01:00
* `coalesce(string1, string2, ...)` - Returns the first non-empty value from
the given arguments. At least two arguments must be provided.
2015-10-11 08:59:21 +02:00
* `compact(list)` - Removes empty string elements from a list. This can be
useful in some cases, for example when passing joined lists as module
variables or when parsing module outputs.
Example: `compact(module.my_asg.load_balancer_names)`
* `concat(list1, list2)` - Combines two or more lists into a single list.
Example: `concat(aws_instance.db.*.tags.Name, aws_instance.web.*.tags.Name)`
2014-08-19 22:16:29 +02:00
* `distinct(list)` - Removes duplicate items from a list. Keeps the first
occurrence of each element, and removes subsequent occurences. This
function is only valid for flat lists. Example: `distinct(var.usernames)`
2015-03-02 18:40:24 +01:00
* `element(list, index)` - Returns a single element from a list
at the given index. If the index is greater than the number of
elements, this function will wrap using a standard mod algorithm.
This function only works on flat lists. Examples:
* `element(aws_subnet.foo.*.id, count.index)`
* `element(var.list_of_strings, 2)`
2015-03-02 18:40:24 +01:00
2014-08-19 22:16:29 +02:00
* `file(path)` - Reads the contents of a file into the string. Variables
in this file are _not_ interpolated. The contents of the file are
read as-is. The `path` is interpreted relative to the working directory.
[Path variables](#path-variables) can be used to reference paths relative
to other base locations. For example, when using `file()` from inside a
module, you generally want to make the path relative to the module base,
like this: `file("${path.module}/file")`.
2014-07-28 19:43:00 +02:00
2015-03-02 19:27:58 +01:00
* `format(format, args...)` - Formats a string according to the given
format. The syntax for the format is standard `sprintf` syntax.
2016-01-14 21:55:39 +01:00
Good documentation for the syntax can be [found here](https://golang.org/pkg/fmt/).
2015-03-02 19:27:58 +01:00
Example to zero-prefix a count, used commonly for naming servers:
`format("web-%03d", count.index + 1)`.
2015-03-02 19:27:58 +01:00
* `formatlist(format, args...)` - Formats each element of a list
according to the given format, similarly to `format`, and returns a list.
Non-list arguments are repeated for each list element.
For example, to convert a list of DNS addresses to a list of URLs, you might use:
`formatlist("https://%s:%s/", aws_instance.foo.*.public_dns, var.port)`.
If multiple args are lists, and they have the same number of elements, then the formatting is applied to the elements of the lists in parallel.
Example:
`formatlist("instance %v has private ip %v", aws_instance.foo.*.id, aws_instance.foo.*.private_ip)`.
Passing lists with different lengths to formatlist results in an error.
* `index(list, elem)` - Finds the index of a given element in a list.
This function only works on flat lists.
Example: `index(aws_instance.foo.*.tags.Name, "foo-test")`
* `join(delim, list)` - Joins the list with the delimiter for a resultant string.
This function works only on flat lists.
Examples:
* `join(",", aws_instance.foo.*.id)`
* `join(",", var.ami_list)`
2014-10-10 06:23:49 +02:00
* `jsonencode(item)` - Returns a JSON-encoded representation of the given
item, which may be a string, list of strings, or map from string to string.
Note that if the item is a string, the return value includes the double
quotes.
* `keys(map)` - Returns a lexically sorted list of the map keys.
* `length(list)` - Returns a number of members in a given list, map, or string.
2015-04-15 19:55:49 +02:00
or a number of characters in a given string.
* `${length(split(",", "a,b,c"))}` = 3
* `${length("a,b,c")}` = 5
* `${length(map("key", "val"))}` = 1
2015-04-15 19:55:49 +02:00
* `list(items...)` - Returns a list consisting of the arguments to the function.
This function provides a way of representing list literals in interpolation.
* `${list("a", "b", "c")}` returns a list of `"a", "b", "c"`.
* `${list()}` returns an empty list.
* `lookup(map, key [, default])` - Performs a dynamic lookup into a mapping
2015-02-20 19:04:53 +01:00
variable. The `map` parameter should be another variable, such
as `var.amis`. If `key` does not exist in `map`, the interpolation will
fail unless you specify a third argument, `default`, which should be a
string value to return if no `key` is found in `map`. This function
only works on flat maps and will return an error for maps that
include nested lists or maps.
* `lower(string)` - Returns a copy of the string with all Unicode letters mapped to their lower case.
* `map(key, value, ...)` - Returns a map consisting of the key/value pairs
specified as arguments. Every odd argument must be a string key, and every
even argument must have the same type as the other values specified.
Duplicate keys are not allowed. Examples:
* `map("hello", "world")`
* `map("us-east", list("a", "b", "c"), "us-west", list("b", "c", "d"))`
* `md5(string)` - Returns a (conventional) hexadecimal representation of the
MD5 hash of the given string.
2015-03-02 18:40:24 +01:00
* `replace(string, search, replace)` - Does a search and replace on the
given string. All instances of `search` are replaced with the value
of `replace`. If `search` is wrapped in forward slashes, it is treated
as a regular expression. If using a regular expression, `replace`
can reference subcaptures in the regular expression by using `$n` where
`n` is the index or name of the subcapture. If using a regular expression,
the syntax conforms to the [re2 regular expression syntax](https://code.google.com/p/re2/wiki/Syntax).
* `sha1(string)` - Returns a (conventional) hexadecimal representation of the
SHA-1 hash of the given string.
Example: `"${sha1("${aws_vpc.default.tags.customer}-s3-bucket")}"`
* `sha256(string)` - Returns a (conventional) hexadecimal representation of the
SHA-256 hash of the given string.
Example: `"${sha256("${aws_vpc.default.tags.customer}-s3-bucket")}"`
* `signum(int)` - Returns -1 for negative numbers, 0 for 0 and 1 for positive numbers.
This function is useful when you need to set a value for the first resource and
a different value for the rest of the resources.
Example: `element(split(",", var.r53_failover_policy), signum(count.index))`
where the 0th index points to `PRIMARY` and 1st to `FAILOVER`
* `sort(list)` - Returns a lexographically sorted list of the strings contained in
the list passed as an argument. Sort may only be used with lists which contain only
strings.
Examples: `sort(aws_instance.foo.*.id)`, `sort(var.list_of_strings)`
2015-03-02 18:40:24 +01:00
* `split(delim, string)` - Splits the string previously created by `join`
back into a list. This is useful for pushing lists through module
outputs since they currently only support string values. Depending on the
use, the string this is being performed within may need to be wrapped
in brackets to indicate that the output is actually a list, e.g.
`a_resource_param = ["${split(",", var.CSV_STRING)}"]`.
2015-03-02 18:40:24 +01:00
Example: `split(",", module.amod.server_ids)`
2016-01-30 10:51:28 +01:00
* `trimspace(string)` - Returns a copy of the string with all leading and trailing white spaces removed.
2016-01-30 00:28:04 +01:00
* `upper(string)` - Returns a copy of the string with all Unicode letters mapped to their upper case.
* `uuid()` - Returns a UUID string in RFC 4122 v4 format. This string will change with every invocation of the function, so in order to prevent diffs on every plan & apply, it must be used with the [`ignore_changes`](/docs/configuration/resources.html#ignore-changes) lifecycle attribute.
* `values(map)` - Returns a list of the map values, in the order of the keys
returned by the `keys` function. This function only works on flat maps and
will return an error for maps that include nested lists or maps.
## Templates
Long strings can be managed using templates. [Templates](/docs/providers/template/index.html) are [resources](/docs/configuration/resources.html) defined by a filename and some variables to use during interpolation. They have a computed `rendered` attribute containing the result.
A template resource looks like:
```
resource "template_file" "example" {
template = "${hello} ${world}!"
vars {
hello = "goodnight"
world = "moon"
}
}
output "rendered" {
value = "${template_file.example.rendered}"
}
```
Then the rendered value would be `goodnight moon!`.
You may use any of the built-in functions in your template.
### Using Templates with Count
Here is an example that combines the capabilities of templates with the interpolation
from `count` to give us a parametized template, unique to each resource instance:
```
variable "count" {
default = 2
}
variable "hostnames" {
default = {
"0" = "example1.org"
"1" = "example2.net"
}
}
resource "template_file" "web_init" {
// here we expand multiple template_files - the same number as we have instances
count = "${var.count}"
template = "${file("templates/web_init.tpl")}"
vars {
// that gives us access to use count.index to do the lookup
hostname = "${lookup(var.hostnames, count.index)}"
}
}
resource "aws_instance" "web" {
// ...
count = "${var.count}"
// here we link each web instance to the proper template_file
user_data = "${element(template_file.web_init.*.rendered, count.index)}"
}
```
With this, we will build a list of `template_file.web_init` resources which we can
use in combination with our list of `aws_instance.web` resources.
## Math
Simple math can be performed in interpolations:
```
variable "count" {
default = 2
}
resource "aws_instance" "web" {
// ...
count = "${var.count}"
// tag the instance with a counter starting at 1, ie. web-001
tags {
Name = "${format("web-%03d", count.index + 1)}"
}
}
```
The supported operations are:
- *Add* (`+`), *Subtract* (`-`), *Multiply* (`*`), and *Divide* (`/`) for **float** types
- *Add* (`+`), *Subtract* (`-`), *Multiply* (`*`), *Divide* (`/`), and *Modulo* (`%`) for **integer** types
-> **Note:** Since Terraform allows hyphens in resource and variable names,
it's best to use spaces between math operators to prevent confusion or unexpected
behavior. For example, `${var.instance-count - 1}` will subtract **1** from the
`instance-count` variable value, while `${var.instance-count-1}` will interpolate
the `instance-count-1` variable value.