EXECUTE IMMEDIATE FROM | Snowflake Documentation (original) (raw)

EXECUTE IMMEDIATE FROM executes the SQL statements specified in a file in a stage. The file can contain SQL statements or Snowflake Scripting blocks. The statements must be syntactically correct SQL statements.

You can use the EXECUTE IMMEDIATE FROM command to execute the statements in a file from any Snowflake session.

This feature provides a mechanism to control the deployment and management of your Snowflake objects and code. For example, you can execute a stored script to create a standard Snowflake environment for all your accounts. The configuration script might include statements that create users, roles, databases, and schemas for every new account.

Jinja2 templating¶

EXECUTE IMMEDIATE FROM can also execute a template file using the Jinja2 templating language. A template can contain variables and expressions, enabling the use of loops, conditionals, variable substitution, macros, and more. Templates can also include other templates and can import macros defined in other files located on a stage.

For more information about the templating language, see the Jinja2 documentation.

The template file to be executed must be:

• A syntactically valid Jinja2 template.
• Located in a stage or Git repository.
• Able to render syntactically valid SQL statements.

Templating enables more flexible control structures and parameterization using environment variables. For example, you can use a template to dynamically choose the deployment target of the objects defined in the script. To use a template to render a SQL script, use the templating directive or add aUSING clause with at least one template variable.

Templating directive¶

You can use either one of the two templating directives.

The recommended directive uses valid SQL syntax:

Optionally, you can use the alternative directive:

Note

Only a byte order mark and up to 10 whitespace characters (newlines, tabs, spaces) may be placed in front of the directive. Any characters that come after the directive on the same line will be ignored.

Using content from staged files in a template¶

A template can load other staged files either directly through theSnowflakeFile APIor through Jinja2’s include,import, andinheritance features.

Files can be referenced by absolute paths:

{% include "@my_stage/path/to/my_template" %} {% import "@my_stage/path/to/my_template" as my_template %} {% extends "@my_stage/path/to/my_template" %} {{ SnowflakeFile.open("@my_stage/path/to/my_template", 'r', require_scoped_url = False).read() }}

Include, import, and extends also support relative paths while the SnowflakeFile API supports scoped Snowflake file URLs:

{% include "my_template" %} {% import "../my_template" as my_template %} {% extends "/path/to/my_template" %}

See also:

EXECUTE IMMEDIATE

Syntax¶

EXECUTE IMMEDIATE FROM { absoluteFilePath | relativeFilePath } [ USING ( => [ , => [ , ... ] ] ) ] [ DRY_RUN = { TRUE | FALSE } ]

Where:

Required parameters¶

Absolute file path (absoluteFilePath)¶

_namespace_

Database and/or schema in which the internal or external stage resides, in the form of _databasename_._schemaname_or _schemaname_. The namespace is optional if a database and schema are currently in use for the user session; otherwise, it is required.

_stagename_

Name of the internal or external stage.

_path_

Case-sensitive path to the file in the stage.

_filename_

Name of the file to execute. It must contain syntactically correct and valid SQL statements. Each statement must be separated by a semicolon.

Relative file path (relativeFilePath)¶

_path_

Case-sensitive relative path to the file in the stage. Relative paths support established conventions such as a leading /to indicate the root of a stage’s file system, ./ to refer to the current directory (the directory the parent file is located in) and ../ to refer to the parent directory. For more information, see Usage notes.

_filename_

Name of the file to execute. It must contain syntactically correct and valid SQL statements. Each statement must be separated by a semicolon.

Optional parameters¶

USING ( <key> => <value> [ , <key> => <value> [ , ... ] ] )

Allows you to pass one or more key-value pairs that can be used to parameterize template expansion. The key-value pairs must form a comma-separated list.

When the USING clause is present, the file is first rendered as a Jinja2 templatebefore being executed as a SQL script.

Where:

• _key_ is the name of the template variable. The template variable name can optionally be enclosed in double quotes (").
• _value_ is the value to assign to the variable in the template. String values must be enclosed in ' or$$. For an example, see Templating usage notes.

DRY_RUN = { TRUE | FALSE }

Specifies whether to preview the rendered file without executing it as a SQL script.

• TRUE returns the rendered file contents without executing the SQL statements.
• FALSE renders SQL statements from the template and executes those statements.

Default: FALSE

Returns¶

EXECUTE IMMEDIATE FROM returns:

• The result of the last statement in the file if all statements are successfully executed.
• The error message, if any statement in the file failed.
  If there is an error in any statement in the file, the EXECUTE IMMEDIATE FROM command fails and returns the error message of the failed statement.
  Note
  If the EXECUTE IMMEDIATE FROM command fails and returns an error message, any statements in the file prior to the failed statement have successfully completed.

Access control requirements¶

• The role used to execute the EXECUTE IMMEDIATE FROM command must have the USAGE (external stage) or READ (internal stage) privilege on the stage where the file is located.
• The role used to execute the file can only execute the statements in the file for which it has privileges. For example, if there is a CREATE TABLE statement in the file, the role must have thenecessary privileges to create a table in the account or the statement fails.

The USAGE privilege on the parent database and schema are required to perform operations on any object in a schema.

For instructions on creating a custom role with a specified set of privileges, see Creating custom roles.

For general information about roles and privilege grants for performing SQL actions onsecurable objects, see Overview of Access Control.

Usage notes¶

• The SQL statements in a file to be executed can include EXECUTE IMMEDIATE FROM statements:
  • Nested EXECUTE IMMEDIATE FROM statements can use relative file paths.
    Relative paths are evaluated in respect to the stage and file path of the parent file. If the relative file path starts with/, the path starts at the root directory of the stage containing the parent file.
    For an example, see Examples.
  • Relative file paths must be enclosed in single quotes (') or $$.
  • The maximum execution depth for nested files is 5.
• Absolute file paths can optionally be enclosed in single quotes (') or $$.
• The file to be executed cannot be larger than 10MB in size.
• The file to be executed must be encoded in UTF-8.
• The file to be executed must be uncompressed. If you use the PUT command to upload a file to an internal stage, you must explicitly set the AUTO_COMPRESS parameter to FALSE.
  For example, upload my_file.sql to my_stage:
  PUT file://~/sql/scripts/my_file.sql @my_stage/scripts/
  AUTO_COMPRESS=FALSE;
• The execution of all files in a directory is not supported. For example, EXECUTE IMMEDIATE FROM @stage_name/scripts/results in an error.

Templating usage notes¶

• Variable names in templates are case-sensitive.
• The template variable name can be optionally enclosed in double quotes. Enclosing the variable name can be useful if anyreserved keywords are used as variable names.
• The following parameter types are supported in the USING clause:
  • String. Must be enclosed by ' or $$. For example, USING (a => 'a', b => <span class="katex"><span class="katex-mathml"><math xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow></mrow><annotation encoding="application/x-tex"></annotation></semantics></math></span><span class="katex-html" aria-hidden="true"></span></span>b$$).
  • Number (decimal and integer). For example, USING (a => 1, b => -1.23).
  • Boolean. For example, USING (a => TRUE, b => FALSE).
  • NULL. For example, USING (a => NULL).
    Note
    The Jinja2 templating engine interprets a NULL value as the Python NoneType type.
  • Session variables. For example, USING (a => $var). Only session variables holding values of supported data types are allowed.
  • Bind variables. For example, USING (a => :var). Only bind variables holding values of supported data types are allowed. You can use bind variables to pass stored procedure arguments to a template.
• Files on repository stages or in Snowflake Native Apps cannot be accessed from the template.
• The maximum result size for template rendering is 100,000 bytes.
• Templates are rendered using the Jinja2 version 3.1.4 templating engine.

Troubleshooting EXECUTE IMMEDIATE FROM errors¶

This section contains some common errors that result from an EXECUTE IMMEDIATE FROM statement and how you can resolve them.

• File errors
• Stage errors
• Access control errors
• Templating errors

File errors¶

Stage errors¶

Access control errors¶

See also: Stage errors.

Templating errors¶

Examples¶

Basic example¶

This example executes the file create-inventory.sql located in stage my_stage.

1. Create a file named create-inventory.sql with the following statements:
   CREATE OR REPLACE TABLE my_inventory(
   sku VARCHAR,
   price NUMBER
   );
   EXECUTE IMMEDIATE FROM './insert-inventory.sql';
   SELECT sku, price
   FROM my_inventory
   ORDER BY price DESC;
2. Create a file named insert-inventory.sql with the following statements:
   INSERT INTO my_inventory
   VALUES ('XYZ12345', 10.00),
   ('XYZ81974', 50.00),
   ('XYZ34985', 30.00),
   ('XYZ15324', 15.00);
3. Create an internal stage my_stage:
4. Upload both local files to the stage using the PUT command:
   PUT file:///sql/scripts/create-inventory.sql @my_stage/scripts/
   AUTO_COMPRESS=FALSE;
   PUT file:///sql/scripts/insert-inventory.sql @my_stage/scripts/
   AUTO_COMPRESS=FALSE;
5. Execute the create-inventory.sql script located in my_stage:
   EXECUTE IMMEDIATE FROM @my_stage/scripts/create-inventory.sql;
   Returns:
   +----------+-------+
   | SKU | PRICE |
   |----------+-------|
   | XYZ81974 | 50 |
   | XYZ34985 | 30 |
   | XYZ15324 | 15 |
   | XYZ12345 | 10 |
   +----------+-------+

A simple template example¶

1. Create a template file setup.sql with two variables and the templating directive:
   --!jinja
   CREATE SCHEMA {{env}};
   CREATE TABLE RAW (COL OBJECT)
   DATA_RETENTION_TIME_IN_DAYS = {{retention_time}};
2. Create a stage — optional if you already have a stage to which you can upload files.
   For example, create an internal stage in Snowflake:
3. Upload the file to your stage.
   For example, use the PUT command from your local environment to upload file setup.sqlto stage my_stage:
   PUT file://path/to/setup.sql @my_stage/scripts/
   AUTO_COMPRESS=FALSE;
4. Execute the file setup.sql:
   EXECUTE IMMEDIATE FROM @my_stage/scripts/setup.sql
   USING (env=>'dev', retention_time=>0);

A template example with macros, conditionals, loops, and imports¶

1. Create a template file containing a macro definition.
   For example, create a file macros.jinja in your local environment:
   {%- macro get_environments(deployment_type) -%}
   {%- if deployment_type == 'prod' -%}
   {{ "prod1,prod2" }}
   {%- else -%}
   {{ "dev,qa,staging" }}
   {%- endif -%}
   {%- endmacro -%}
2. Create a template file and add the templating directive (--!jinja2) to the top of the file.
   After the templating directive, add an import statement to import the macro defined in the file that you created in the previous step. For example, create a file setup-env.sql in your local environment:
   --!jinja2
   {% from "macros.jinja" import get_environments %}
   {%- set environments = get_environments(DEPLOYMENT_TYPE).split(",") -%}
   {%- for environment in environments -%}
   CREATE DATABASE {{ environment }}_db;
   USE DATABASE {{ environment }}_db;
   CREATE TABLE {{ environment }}_orders (
   id NUMBER,
   item VARCHAR,
   quantity NUMBER);
   CREATE TABLE {{ environment }}_customers (
   id NUMBER,
   name VARCHAR);
   {% endfor %}
3. Create a stage — optional if you already have a stage to which you can upload files.
   For example, create an internal stage in Snowflake:
4. Upload the file to your stage.
   For example, use the PUT command from your local environment to upload the filessetup-env.sql and macros.jinja to the stage my_stage:
   PUT file://path/to/setup-env.sql @my_stage/scripts/
   AUTO_COMPRESS=FALSE;
   PUT file://path/to/macros.jinja @my_stage/scripts/
   AUTO_COMPRESS=FALSE;
5. Preview the SQL statements rendered by the template to check for any problems with your Jinja2 code:
   EXECUTE IMMEDIATE FROM @my_stage/scripts/setup-env.sql
   USING (DEPLOYMENT_TYPE => 'prod') DRY_RUN = TRUE;
   Returns:
   +----------------------------------+

   rendered file contents
   --!jinja2
   CREATE DATABASE prod1_db;
   USE DATABASE prod1_db;
   CREATE TABLE prod1_orders (
   id NUMBER,
   item VARCHAR,
   quantity NUMBER);
   CREATE TABLE prod1_customers (
   id NUMBER,
   name VARCHAR);
   CREATE DATABASE prod2_db;
   USE DATABASE prod2_db;
   CREATE TABLE prod2_orders (
   id NUMBER,
   item VARCHAR,
   quantity NUMBER);
   CREATE TABLE prod2_customers (
   id NUMBER,
   name VARCHAR);

+----------------------------------+ 6. Execute the file setup-env.sql:
EXECUTE IMMEDIATE FROM @my_stage/scripts/setup-env.sql
USING (DEPLOYMENT_TYPE => 'prod');