ドキュメント変更履歴 2026-08-20

--- ahttps://docs.snowflake.com/en/developer-guide/snowpark/python/checkpoints-validators+++ bhttps://docs.snowflake.com/en/developer-guide/snowpark/python/checkpoints-validators@@ -27,17 +27,17 @@    ```python-   def original_spark_code_I_dont_understand(df):-   from pyspark.sql.functions import col, when--   ret = df.withColumn(-    "life_stage",-    when(col("byte") < 4, "child")-    .when(col("byte").between(4, 10), "teenager")-    .otherwise("adult"),-   )-   return ret+  def original_spark_code_I_dont_understand(df):+    from pyspark.sql.functions import col, when++    ret = df.withColumn(+        "life_stage",+        when(col("byte") < 4, "child")+        .when(col("byte").between(4, 10), "teenager")+        .otherwise("adult"),+    )+    return ret    @check_with_spark(-   job_context=job_context, spark_function=original_spark_code_I_dont_understand+    job_context=job_context, spark_function=original_spark_code_I_dont_understand   )   def new_snowpark_code_I_do_understand(df):@@ -45,15 +45,15 @@      ref = df.with_column(-     "life_stage",-     when(col("byte") < 4, lit("child"))-     .when(col("byte").between(4, 10), lit("teenager"))-     .otherwise(lit("adult")),-   )-   return ref--   df1 = new_snowpark_code_I_do_understand(df)+        "life_stage",+        when(col("byte") < 4, lit("child"))+        .when(col("byte").between(4, 10), lit("teenager"))+        .otherwise(lit("adult")),+    )+    return ref++  df1 = new_snowpark_code_I_do_understand(df)   ``` -- <span className="emph">validate_dataframe_checkpoint:</span> This function validates a Snowpark Dataframe against a specific checkpoint schema file or imported Dataframe according to the argument mode. It ensures that the information collected for that DataFrame and the DataFrame that is passed to the function are equivalent.+- <span className="emph">validate_dataframe_checkpoint:</span> This function validates a Snowpark DataFrame against a specific checkpoint schema file or imported DataFrame according to the argument mode. It ensures that the information collected for that DataFrame and the DataFrame that is passed to the function are equivalent.  @@ -64,6 +64,6 @@   - `df` (SnowparkDataFrame): The DataFrame to validate   - `checkpoint_name` (str): The name of the checkpoint to validate against-  - `job_context` (SnowparkJobContext, optional) (str): The job context for the validation; required for PARQUET mode-  - `mode` (CheckpointMode): The mode of validation (e.g., SCHEMA, PARQUET); defaults to SCHEMA+  - `job_context` (SnowparkJobContext, optional): The job context for the validation; required for DATAFRAME mode+  - `mode` (CheckpointMode): The mode of validation (for example, SCHEMA or DATAFRAME); defaults to SCHEMA   - `custom_checks` (Optional[dict[Any, Any]], optional): Custom checks to apply during validation   - `skip_checks` (Optional[dict[Any, Any]], optional): Checks to skip during validation@@ -88,7 +88,7 @@   ``` -  Depending on the mode selected, the validation will use either the collected schema file or a Parquet-loaded Dataframe in Snowflake to verify the equivalence against the PySpark version.--- <span className="emph">check-output_schema:</span> This decorator validates the schema of a Snowpark function's output and ensures that the output DataFrame conforms to a specified Pandera schema. It is particularly useful for enforcing data integrity and consistency in Snowpark pipelines. This decorator takes several parameters, including the Pandera schema to validate against, the checkpoint name, sampling parameters, and an optional job context. It wraps the Snowpark function and performs schema validation on the output DataFrame before returning the result.+  Depending on the mode selected, the validation will use either the collected schema file or a DataFrame loaded in Snowflake to verify the equivalence against the PySpark version.++- <span className="emph">check_output_schema:</span> This decorator validates the schema of a Snowpark function's output and ensures that the output DataFrame conforms to a specified Pandera schema. It is particularly useful for enforcing data integrity and consistency in Snowpark pipelines. This decorator takes several parameters, including the Pandera schema to validate against, the checkpoint name, sampling parameters, and an optional job context. It wraps the Snowpark function and performs schema validation on the output DataFrame before returning the result.    Example:@@ -156,7 +156,7 @@   @check_input_schema(input_schema, "input_schema_checkpoint")   def process_dataframe(dataframe: SnowparkDataFrame):-  return dataframe.with_column(-   "COLUMN3", dataframe["COLUMN1"] + dataframe["COLUMN2"]-  )+    return dataframe.with_column(+        "COLUMN3", dataframe["COLUMN1"] + dataframe["COLUMN2"]+    )    # Create a Snowpark session and DataFrame@@ -176,5 +176,5 @@ ## Statistics checks -Statistics validations are applied to the specific column type by default when the validation is run in `Schema` mode; these checks can be skipped with `skip_checks`.+Statistics validations are applied to the specific column type by default when the validation is run in `SCHEMA` mode; these checks can be skipped with `skip_checks`.  <div className="colwidths-given">@@ -237,5 +237,5 @@ - `str_startswith` - `in_range`-- `​​equal_to`+- `equal_to` - `greater_than_or_equal_to` - `greater_than` 

--- ahttps://docs.snowflake.com/en/user-guide/cortex-code/cli-reference+++ bhttps://docs.snowflake.com/en/user-guide/cortex-code/cli-reference@@ -6,5 +6,13 @@ - [CoCo in Snowsight](/user-guide/cortex-code/cortex-code-snowsight) -Command line reference for CoCo CLI.+Command line reference for CoCo CLI. CoCo CLI operates in two modes:++- **Interactive mode** (default): Start a persistent session where you type prompts and slash commands at an+  interactive prompt. CoCo maintains conversation context across multiple exchanges.+- **Batch mode**: Pass a single prompt on the command line or from a file. CoCo processes it and exits+  without starting an interactive session.++The sections below cover the startup command, CLI options shared by both modes, subcommands you run+*before* entering a session, interactive-mode commands you use *within* a session, and batch-mode usage.  ## Starting CoCo@@ -65,4 +73,7 @@ ## Commands +These subcommands run from your shell *before* or *instead of* starting an interactive session.+They are distinct from the slash commands (like `/mcp`) that you type inside an interactive session.+ ### `update` @@ -109,5 +120,14 @@ ## Interactive mode +Interactive mode is the default when you run `cortex` without `-p` or `exec`. CoCo opens a persistent+session where you type natural-language prompts and slash commands at the `>` prompt. The agent processes+your request, potentially using multiple tools, and then returns to the prompt for your next input.++Keyboard shortcuts and slash commands described in this section work only in interactive mode. They+don't apply in batch mode or inside a file passed to `cortex exec --file`.+ ### Keyboard shortcuts++These shortcuts are active at the interactive prompt and while the agent is working:  <div className="colwidths-given">@@ -149,4 +169,8 @@  ### Slash commands++Slash commands are special commands you type at the interactive prompt. They start with `/` and+control CoCo's behavior without being sent to the AI model as a prompt. Slash commands work only+in interactive mode.  #### Session management@@ -261,4 +285,8 @@ #### `/worktree`: Git worktrees +Git worktrees let you work on multiple branches simultaneously without switching back and forth.+CoCo creates worktrees under your project directory by default (in a `.worktrees/` subdirectory+or alongside your main checkout, depending on your git configuration).+ | Command                           | Description         | | --------------------------------- | ------------------- |@@ -269,4 +297,9 @@  #### `/sandbox`: Sandbox control++CoCo's sandbox isolates shell commands in a restricted environment so they can't modify files+outside your working directory or access sensitive system resources. This is a CoCo-specific+security feature. The container sandbox uses Docker, while the runtime sandbox uses OS-level+restrictions (macOS sandbox or Linux namespaces).  | Command                 | Description                   |@@ -288,4 +321,11 @@  ## Batch mode++Batch mode runs a single prompt without starting an interactive session. CoCo processes the request,+prints the result, and exits. This is useful for scripting, CI/CD pipelines, and one-off questions+where you don't need a back-and-forth conversation.++Use `-p` for inline prompts or `cortex exec` for prompts from files or stdin. Slash commands don't+work in batch mode because there's no interactive prompt.  | Command                                            | Description                | 

--- ahttps://docs.snowflake.com/en/migrations/aim-for-datawarehouses/manual-migration/README+++ bhttps://docs.snowflake.com/en/migrations/aim-for-datawarehouses/manual-migration/README@@ -1,3 +1,3 @@-# When you want to orchesrate the migration+# When you want to orchestrate the migration Snowflake's recommended way of migrating data warehouses is through the [%migration-skill%](/migrations/aim-for-datawarehouses/overview). Its CoCo-powered orchestration greatly accelerates the overall time it takes to complete a migration. @@ -8,19 +8,19 @@ # What is %scai%? -With the %scai% , migration engineers can:+With the %scai%, migration engineers can:  * extract code from their source platform * run a deterministic conversion on that code- * further advance their migration using AI conversion to cover objects that the deterministic engine could not translate+* further advance their migration using AI conversion to cover objects that the deterministic engine could not translate * deploy that code to Snowflake * migrate data from the source system to Snowflake * validate that data between the two systems -The %scai%  will also allow developers to create skills and agents that utilize the tool to automate their process.+The %scai% will also allow developers to create skills and agents that utilize the tool to automate their process.  ## Prerequisites  * macOS, Windows, or Linux-* SnowflakeCLI: recommended for Snowflake connection configuration [SnowCLI Install Guide](https://docs.snowflake.com/en/developer-guide/snowflake-cli/installation/installation)+* Snowflake CLI: recommended for Snowflake connection configuration [SnowCLI Install Guide](https://docs.snowflake.com/en/developer-guide/snowflake-cli/installation/installation)  * A source database to extract from, or a set of code to use@@ -28,5 +28,5 @@ ## Snowflake Connection Setup -The %scai%  reuses your Snowflake CLI connection configuration. The connection is used for functionality in ai-convert, deploy, and the cloud versions of data migration and validation. Your Snowflake account also authenticates you for %scai% , skipping the need for an access code that was necessary in prior versions.\*+The %scai% reuses your Snowflake CLI connection configuration. The connection is used for functionality in ai-convert, deploy, and the cloud versions of data migration and validation. Your Snowflake account also authenticates you for %scai%, skipping the need for an access code that was necessary in prior versions.  **Snowflake Account Requirement**@@ -34,8 +34,8 @@ Before using scai init, scai code convert, scai code extract, scai ai-convert, or scai code deploy, ensure that you: -Can connect to Snowflake with snow connection test-Your Snowflake CLI has a default connection configured (this is used when you don't specify a name).--`To configure a Snowflake connection:`+- Can connect to Snowflake with `snow connection test`+- Have a default Snowflake CLI connection configured (this is used when you don't specify a name)++To configure a Snowflake connection:  ```shell@@ -71,5 +71,5 @@ If you do not have homebrew installed, follow the [instructions here](https://brew.sh/). -There are two public channels for builds, Preview and GA.+There are two public channels for builds: Preview and GA.  Stable Version (recommended)@@ -93,5 +93,5 @@ Usage -After installation, you can use the %scai% :+After installation, you can use the %scai%:  ```shell@@ -101,5 +101,5 @@ Managing Installations -View installed Version+View installed version  ```shell@@ -168,5 +168,5 @@ ## Accept Terms and Conditions -Issuing the following command will display the license terms for using the %scai% .  It is required that you do this in order to use the product.+Issuing the following command will display the license terms for using the %scai%. It is required that you do this in order to use the product.  ```shell@@ -334,5 +334,5 @@ The interactive mode will prompt you for connection details. -Set a default source connection (used with scai code extract runs without -source-connection)+Set a default source connection (used with scai code extract runs without --source-connection)  ```shell@@ -927,5 +927,5 @@ scai code convert -# Review converted code in converted/ folder, then deploy+# Review converted code in snowflake/ folder, then deploy scai code deploy --all ```@@ -1024,7 +1024,7 @@ "Connection not found" (source database) -1. List your connections: scai connection list -l \&lt;language\&gt;-2. Add a connection if needed: scai connection add-sql-server or scai connection add-redshift-3. Or set a default: scai connection set-default -l \&lt;language\&gt; --connection-name \&lt;name\&gt;+1. List your connections: `scai connection list -l <language>`+2. Add a connection if needed: `scai connection add-sql-server` or `scai connection add-redshift`+3. Or set a default: `scai connection set-default -l <language> --connection-name <name>`  "Authentication failed" for Snowflake 

--- ahttps://docs.snowflake.com/en/release-notes/preview-features+++ bhttps://docs.snowflake.com/en/release-notes/preview-features@@ -79,4 +79,25 @@   <tbody>     <tr>+      <td>Snowflake ODBC Driver 4.x, built on the Universal Core</td>+      <td>Open</td>+      <td>August 2026</td>+      <td>[](/developer-guide/odbc/odbc-universal-core)</td>+      <td>A new major version of the ODBC driver, built on the Universal Core shared Rust library. Installing it replaces the 3.x driver on the same machine, so validate it on a dedicated host, VM, or container. Certificate revocation checking uses certificate revocation lists (CRLs) rather than the Online Certificate Status Protocol (OCSP), and is off by default.</td>+    </tr>+    <tr>+      <td>Snowflake Connector for Python 5.x, built on the Universal Core</td>+      <td>Open</td>+      <td>August 2026</td>+      <td>[](/developer-guide/python-connector/python-connector-universal-core)</td>+      <td>A new major version of the Python connector, built on the Universal Core shared Rust library and published as a release candidate under the existing `snowflake-connector-python` package name. Because pre-release versions are not installed by default, `pip install --pre` is required. Certificate revocation checking uses CRLs rather than OCSP, and is off by default.</td>+    </tr>+    <tr>+      <td>Access externally managed %iceberg-tm% tables through Snowflake Horizon Catalog</td>+      <td>Open</td>+      <td>August 2026</td>+      <td>[](/user-guide/externally-managed-iceberg-tables-access-horizon-irc)</td>+      <td>Use the Horizon Iceberg REST Catalog API to read and perform DML and DDL operations on externally managed Iceberg tables in a catalog-linked database from external engines such as Apache Spark™, Trino, DuckDB, or PyIceberg.</td>+    </tr>+    <tr>       <td>Open Data Sharing</td>       <td>Open</td>@@ -154,11 +175,4 @@       <td>[](/user-guide/classify-intro#label-classify-ai-mode)</td>       <td>Use LLMs to identify additional semantic categories beyond standard classification when you enable AI mode on a classification profile in the Trust Center or with SQL.</td>-    </tr>-    <tr>-      <td>Metadata-based sensitive data classification recommendations in Snowsight</td>-      <td>Open</td>-      <td>July 2026</td>-      <td>[](/user-guide/classify-ui-trust-center#label-classify-trust-center-sensitive-data-banner)</td>-      <td>Show a banner on the database details page in Snowsight when object metadata indicates a database is likely to contain sensitive data, with a one-click option to enable classification.</td>     </tr>     <tr>@@ -419,11 +433,4 @@       <td>[](/user-guide/dcm-projects/dcm-projects-supported-entities#label-dcm-projects-attach-tag)</td>       <td>Declaratively assign Snowflake object tags to %dcm%-managed entities; %dcm% reconciles declared tag assignments on every deployment.</td>-    </tr>-    <tr>-      <td>Power BI ingestion for Semantic View Autopilot</td>-      <td>Open</td>-      <td>June 2026</td>-      <td>[](/user-guide/views-semantic/power-bi-ingestion)</td>-      <td></td>     </tr>     <tr> 

--- ahttps://docs.snowflake.com/en/user-guide/data-integration/openflow/version-history+++ bhttps://docs.snowflake.com/en/user-guide/data-integration/openflow/version-history@@ -254,4 +254,37 @@ </div> +## August 18, 2026++<a id="label-runtime-server-2026-8-18-9"></a>++### Runtime Server 2026.8.18.9++- Security patches and dependency upgrades.+- Upgraded the Runtime UI to version 0.86.0.+  - Prompted users to save or discard unsaved edits before navigating to an inherited Parameter Context, preventing silent loss of configuration changes.++<a id="label-runtime-extensions-2026-8-18-9"></a>++### Runtime Extensions 2026.8.18.9++- Databases: Added a `CountTableRows` processor that counts rows in a SQL table and stores the result in a configured FlowFile attribute.+- Salesforce: Selected the latest staged snapshot for each Salesforce record during deduplication, preventing stale values from being merged when staging contains multiple snapshots.+- CDC SQL Server: Fixed a rare data-loss window where a captured batch could be silently skipped if the destination commit failed after the batch was read.+- Snowflake: Added an opt-in stream-existence cache to `UpdateSnowflakeStream`, avoiding a redundant DDL statement per FlowFile when using `Create Stream If Not Exists`.++<a id="label-connectors-2026-8-18-1"></a>++### Connectors 2026.8.18.1++- MongoDB 0.27.0: Used a 60-second Snowpipe Streaming offset-tracking timeout to reduce transient polling failures while data continues flowing.+- CDC MySQL 0.57.0: Used a 60-second Snowpipe Streaming offset-tracking timeout to reduce transient polling failures while data continues flowing.+- Oracle Embedded License 0.45.0: Increased the default snapshot fetch size from 100 to 5,000 rows and used a 60-second offset-tracking timeout.+- Oracle Embedded License Public Sector 0.44.0: Increased the default snapshot fetch size from 100 to 5,000 rows and used a 60-second offset-tracking timeout.+- Oracle Independent License 0.44.0: Increased the default snapshot fetch size from 100 to 5,000 rows and used a 60-second offset-tracking timeout.+- CDC PostgreSQL 0.62.0: Used a 60-second Snowpipe Streaming offset-tracking timeout to reduce transient polling failures while data continues flowing.+- CDC SQL Server 0.48.0: Used a 60-second Snowpipe Streaming offset-tracking timeout to reduce transient polling failures while data continues flowing.+- SQL Server CT 0.49.0: Added optional connector-event publishing for table consolidation and used a 60-second offset-tracking timeout.+- Table Consolidation 0.1.0: Added a flow that consumes connector events and merges snapshot and CDC rows into consolidated destination tables.+ ## August 13, 2026  

--- ahttps://docs.snowflake.com/en/sql-reference/account-usage+++ bhttps://docs.snowflake.com/en/sql-reference/account-usage@@ -524,4 +524,32 @@       <td>Enterprise Edition (or higher)</td>       <td></td>+    </tr>+    <tr>+      <td>[DATA_MOVEMENT_POLICIES](/sql-reference/account-usage/data_movement_policies)</td>+      <td>Object</td>+      <td>2 hours</td>+      <td></td>+      <td></td>+    </tr>+    <tr>+      <td>[DATA_MOVEMENT_POLICY_RULES](/sql-reference/account-usage/data_movement_policy_rules)</td>+      <td>Object</td>+      <td>2 hours</td>+      <td></td>+      <td></td>+    </tr>+    <tr>+      <td>[DATA_MOVEMENT_RULE_REFERENCES](/sql-reference/account-usage/data_movement_rule_references)</td>+      <td>Historical</td>+      <td>3 hours</td>+      <td></td>+      <td></td>+    </tr>+    <tr>+      <td>[DATA_MOVEMENT_VIOLATIONS](/sql-reference/account-usage/data_movement_violations)</td>+      <td>Historical</td>+      <td>3 hours</td>+      <td></td>+      <td>Data is retained for one year.</td>     </tr>     <tr> 

--- ahttps://docs.snowflake.com/en/developer-guide/snowpark/java/working-with-dataframes+++ bhttps://docs.snowflake.com/en/developer-guide/snowpark/java/working-with-dataframes@@ -163,5 +163,5 @@   ```java   // Create a DataFrame that contains the id, name, and serial_number-  // columns in te "sample_product_data" table.+  // columns in the "sample_product_data" table.   DataFrame df = session.table("sample_product_data").select(     Functions.col("id"), Functions.col("name"), Functions.col("serial_number"));@@ -265,5 +265,5 @@  If the name does not conform to the identifier requirements, you must use double quotes (`"`) around the name. Use a backslash-(`\`) to escape the double quote character within a Scala string literal. For example, the following table name does not start+(`\`) to escape the double quote character within a Java string literal. For example, the following table name does not start with a letter or an underscore, so you must use double quotes around the name: @@ -273,5 +273,5 @@  Note that when specifying the name of a <span className="emph">column</span>, you don't need to use double quotes around the name. The Snowpark library-automatically encloses the column name in double quotes for you if the name does not comply with the identifier requirements:.+automatically encloses the column name in double quotes for you if the name does not comply with the identifier requirements:  ```java@@ -790,5 +790,5 @@ #### Performing a Self-Join -If you need to join a table with itself on different columns, you cannot perform the self-join with a single DataFrame. The+If you need to join a table with itself on different columns, you can't perform the self-join with a single DataFrame. The following examples that use a single DataFrame to perform a self-join fail because the column expressions for `"id"` are present in the left and right sides of the join:@@ -1293,5 +1293,5 @@ ### Merging Rows into a Table -To insert, update, and deletes rows in one table based on values in a second table or a subquery (the equivalent of the+To insert, update, and delete rows in one table based on values in a second table or a subquery (the equivalent of the [](/sql-reference/sql/merge) command in SQL), do the following: @@ -1305,5 +1305,5 @@    - To specify the update or deletion that should be performed on matching rows, call the `whenMatched` method. -     If you need to specify an additional condition whe rows should be updated or deleted, you can pass in a column expression for+     If you need to specify an additional condition when rows should be updated or deleted, you can pass in a column expression for      that condition. @@ -1338,5 +1338,5 @@ ```java MergeResult mergeResult = target.merge(source, target.col("id").equal_to(source.col("id")))-                    .whenNotMatched().insert([source.col("id"), source.col("value")])+                    .whenNotMatched().insert(new Column[]{source.col("id"), source.col("value")})                     .collect(); ```@@ -1483,5 +1483,5 @@ for the table `sample_product_data`: -```scala+```java HasCachedResult dfTempTable = dfTable.cacheResult(); ```@@ -1550,5 +1550,5 @@      // Upload the CSV files in /tmp with names that start with "file".      // You can use the wildcard characters "*" and "?" to match multiple files.-     PutResult[] putResults = session.file().put("file:///tmp/file*.csv", "@myStage/prefix2")+     PutResult[] putResults = session.file().put("file:///tmp/file*.csv", "@myStage/prefix2");      ``` @@ -1862,5 +1862,5 @@    - [PARTITION BY or HEADER](#label-copy-into-location-optional-parameters). -   Note that you cannot use the `option` method to set the following options:+   Note that you can't use the `option` method to set the following options:     - The TYPE format type option. 

--- ahttps://docs.snowflake.com/en/sql-reference/stored-procedures/system_classify_schema+++ bhttps://docs.snowflake.com/en/sql-reference/stored-procedures/system_classify_schema@@ -71,4 +71,22 @@ Classify the specified object while specifying the number of rows to sample and set the recommended system tag on each column in the specified object when the classification process is complete.++When you use this argument, call the stored procedure with the role that has the OWNERSHIP privilege on the schema.++</dd>+<dt>`{'ai_mode': true}`</dt>+<dd>++Enables [AI mode](/user-guide/classify-intro#label-classify-ai-mode) to use LLMs to identify additional semantic categories beyond those+identified by standard classification.++LLM usage is billed under the `AI_SENSITIVE_DATA_CLASSIFICATION` [service type](/sql-reference/service-types).++</dd>+<dt>`{'ai_mode': true, 'auto_tag': true}`</dt>+<dd>++Classify the specified object with AI mode enabled and set the recommended classification system tags on the columns when the+classification process is complete.  When you use this argument, call the stored procedure with the role that has the OWNERSHIP privilege on the schema. 

--- ahttps://docs.snowflake.com/en/user-guide/security-access-control-privileges+++ bhttps://docs.snowflake.com/en/user-guide/security-access-control-privileges@@ -68,4 +68,9 @@     </tr>     <tr>+      <td>APPLY DATA MOVEMENT POLICY</td>+      <td>Global</td>+      <td>Grants the ability to set a [data movement policy](/user-guide/data-movement-policies) on a tag or on the account.</td>+    </tr>+    <tr>       <td>APPLY FEATURE POLICY</td>       <td>Global</td>@@ -555,4 +560,9 @@       <td>APPLY CONTACT</td>       <td>Grants the ability to associate or detach a [contact](/user-guide/contacts-using) with an object.</td>+      <td></td>+    </tr>+    <tr>+      <td>APPLY DATA MOVEMENT POLICY</td>+      <td>Grants the ability to set a [data movement policy](/user-guide/data-movement-policies) on a tag or on the account.</td>       <td></td>     </tr>@@ -1753,4 +1763,12 @@       <td>CREATE DATA METRIC FUNCTION</td>       <td>Enables creating a new data metric function in a schema.</td>+    </tr>+    <tr>+      <td>CREATE DATA MOVEMENT POLICY</td>+      <td>Enables creating a new [data movement policy](/user-guide/data-movement-policies) in a schema.</td>+    </tr>+    <tr>+      <td>CREATE DATA MOVEMENT RULE</td>+      <td>Enables creating a new [data movement rule](/user-guide/data-movement-policies) in a schema.</td>     </tr>     <tr> 

--- ahttps://docs.snowflake.com/en/user-guide/cortex-code/managed-settings+++ bhttps://docs.snowflake.com/en/user-guide/cortex-code/managed-settings@@ -9,5 +9,5 @@ Managed settings allow IT administrators to enforce CoCo behavior across an organization. They apply to both **CoCo CLI** and **CoCo Desktop**, which read the same `managed-settings.json` file from the same system path. Settings are deployed to a system-owned file that users can't modify, and they take precedence over all user-level configuration. -CoCo Desktop and CoCo CLI share this schema, but some keys apply to the CLI only and have no effect in Desktop: `required.minimumVersion` (Desktop has its own auto-update), `settings.forceSandboxEnabled` / `settings.forceSandboxMode`, `files.connectionsFile` / `files.mcpFile`, and `defaults.*`. The `permissions`, `settings.forceNoHistoryMode`, and `ui.*` keys apply to both.+CoCo Desktop and CoCo CLI share this schema, but some keys apply to the CLI only and have no effect in Desktop: `required.minimumVersion` (Desktop has its own auto-update), `settings.forceSandboxEnabled` / `settings.forceSandboxMode`, `files.connectionsFile` / `files.mcpFile`, `defaults.*`, and `account()` (Desktop reads the setting but doesn't enforce it; account restrictions are CLI-only). The `permissions`, `settings.forceNoHistoryMode`, and `ui.*` keys apply to both.  When managed settings are present, CoCo displays a banner at startup indicating that it's running in managed mode.@@ -411,4 +411,12 @@     </tr>     <tr>+      <td>`"grep"`</td>+      <td>The `grep` (file content search) tool</td>+    </tr>+    <tr>+      <td>`"glob"`</td>+      <td>The `glob` (file pattern matching) tool</td>+    </tr>+    <tr>       <td>`"bash"`</td>       <td>All bash shell invocations</td>@@ -476,9 +484,9 @@     <tr>       <td>`"account(myorg)"`</td>-      <td>Exact Snowflake account name</td>+      <td>Exact Snowflake account name (CLI-only; Desktop reads but doesn't enforce this setting)</td>     </tr>     <tr>       <td>`"account(myorg-*)"`</td>-      <td>Snowflake accounts matching a glob</td>+      <td>Snowflake accounts matching a glob (CLI-only; Desktop reads but doesn't enforce this setting)</td>     </tr>   </tbody>@@ -562,4 +570,6 @@       "write",       "edit",+      "grep",+      "glob",       "sql_execute",       "skill(bundled:*)",@@ -637,2 +647,4 @@  To check the enforcement config path for your platform, verify whether the file exists at the OS-specific path listed in [](#label-cortex-code-managed-file-locations).++To inspect the active policy from the CLI, run `cortex managed-settings`. The command prints the parsed schema, the resolved permission mode, and any validation errors or warnings. 

--- ahttps://docs.snowflake.com/en/sql-reference/sql/desc-feature-policy+++ bhttps://docs.snowflake.com/en/sql-reference/sql/desc-feature-policy@@ -3,7 +3,5 @@ This feature is not available in the People's Republic of China. -Available to all accounts.--Describes the properties of a [feature policy](/developer-guide/native-apps/ui-consumer-feature-policies).+Describes the properties of a [feature policy](/user-guide/feature-policies).  DESCRIBE can be abbreviated to DESC.@@ -108,4 +106,11 @@       <td>The list of objects that the feature policy blocks for creation.</td>     </tr>+    <tr>+      <td>`policy_definition`</td>+      <td>+        The YAML body for [conditional creation rules](/user-guide/feature-policies#label-feature-policy-rules),+        if one is set on the policy. Empty when the policy has no body.+      </td>+    </tr>   </tbody>  

--- ahttps://docs.snowflake.com/en/migrations/aim-for-datawarehouses/data-migration-validation/glossary+++ bhttps://docs.snowflake.com/en/migrations/aim-for-datawarehouses/data-migration-validation/glossary@@ -23,11 +23,11 @@ The L3 step that compares individual column values when row fingerprinting detects a mismatch. Results are stored in `CELL_VALIDATION_RESULTS`. +## Compute pool++Snowflake object that hosts Snowpark Container Services. You create a compute pool before deploying Orchestrator or Worker services. See [Deploying workers](./deploy-workers).+ ## Custom metrics  Workflow-wide overrides for L2 aggregate metrics per data type. Set `validationCustomMetrics` at the workflow root with `source` and `target` arrays. See [Customizing normalization and metrics](../manual-migration/data-validation-configuration-reference#customizing-normalization-and-metrics).--## Compute pool--Snowflake object that hosts Snowpark Container Services. You create a compute pool before deploying Orchestrator or Worker services. See [Deploying workers](./deploy-workers).  ## DUPLICATE_SOURCE, DUPLICATE_TARGET, DUPLICATE_BOTH_SIDES@@ -65,5 +65,5 @@ ## NOT_FOUND_SOURCE, NOT_FOUND_TARGET -L3 row validation results for rows present on only one side: `NOT_FOUND_TARGET` (the row exists on the source but is missing from the target) and `NOT_FOUND_SOURCE` (the row exists on the target but is missing from the source). See [Validation levels and result codes](../manual-migration/data-validation-configuration-reference#validation-levels-and-result-codes).+L3 row validation results for rows present on only one side: `NOT_FOUND_SOURCE` (the row exists on the target but is missing from the source) and `NOT_FOUND_TARGET` (the row exists on the source but is missing from the target). See [Validation levels and result codes](../manual-migration/data-validation-configuration-reference#validation-levels-and-result-codes).  ## Orchestrator 

--- ahttps://docs.snowflake.com/en/user-guide/cleanrooms/spec-code-spec+++ bhttps://docs.snowflake.com/en/user-guide/cleanrooms/spec-code-spec@@ -34,5 +34,5 @@     content_hash: <sha256_hash>      # Optional: Lowercase SHA-256 hash for integrity verification -functions:                      # Required if no procedures defined+functions:                      # Required if no procedures or ML jobs defined   - name: <identifier>          # One or more functions...     type: UDF | UDTF            # Required: Function type@@ -55,5 +55,5 @@   - name: <identifier>          # One or more procedure items...     language: PYTHON            # Required: Currently only PYTHON supported-    runtime_version: <python_version>  # Optional: Python runtime version+    runtime_version: <python_version>  # Optional: Python runtime version (3.10 - 3.14)     handler: <handler>          # Required: Handler function     arguments:                  # Optional: One or more procedure arguments@@ -125,6 +125,5 @@ An alias for referencing this artifact in imports. When referencing this alias within this spec, use the bare alias name rather than -<code className="samp">cleanroom.<em>spec_name</em>$<em>alias</em></code>; that is, use the bare function name to reference another function in this-spec.+<code className="samp">cleanroom.<em>spec_name</em>$<em>alias</em></code>.  </dd>@@ -138,5 +137,5 @@  - **The stage must be internal.** External stages aren't supported.-- **The stage must have DIRECTORY enabled**: The stage containing artifacts must have `DIRECTORY = TRUE` set.+- **The stage must have DIRECTORY enabled**: The stage containing artifacts must have `DIRECTORY = (ENABLE = TRUE)` set. - **Stage path format**: Must follow <code className="samp">@[<em>DB</em>.]SCHEMA.STAGE/path/to/file.ext</code> format. - **No path traversal**: Stage paths can't contain `..` or `\`. 

--- ahttps://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-agents-mcp-connectors+++ bhttps://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-agents-mcp-connectors@@ -138,5 +138,5 @@ **GitHub:** -1. Sign into GitHub.+1. Sign in to GitHub. 2. Navigate to the top right avatar, and select **Settings**. 3. Select **Developer Settings**.@@ -210,5 +210,5 @@ 13. Select **Salesforce**. 14. Enter a description for the MCP Server.-15. Enter the MCP server URL. Salesforce exposes their MCP servers through individual URLs, and the user needs to know what server/tools to access. The OAuth tokens retrieved from the last step can be used for all servers and tools, as long as the scopes permit.+15. Enter the MCP server URL. Salesforce exposes its MCP servers through individual URLs, and the user needs to know what server/tools to access. The OAuth tokens retrieved from the last step can be used for all servers and tools, as long as the scopes permit.  The base URL for each hosted MCP server is:@@ -534,6 +534,6 @@   connectors list. Snowflake supports user authentication types for Atlassian, GitHub, Glean, Linear, and Salesforce in the API integration. - **Least-privilege access:** Grant only the minimum required privileges for each role. Access to an MCP-  server doesn't automatically grant access to its tools, you must grant permissions separately for each tool.-- **OAuth authentication:** Snowflake only supports using OAuth for all MCP server connections.+  server doesn't automatically grant access to its tools. You must grant permissions separately for each tool.+- **OAuth authentication:** Snowflake supports only OAuth for MCP server connections. - **Refresh token validity:** When using `TYPE = OAUTH2`, set `OAUTH_REFRESH_TOKEN_VALIDITY`   (in seconds, minimum `3600`) in the API integration's `API_USER_AUTHENTICATION` block. 

--- ahttps://docs.snowflake.com/en/developer-guide/snowpark/scala/working-with-dataframes+++ bhttps://docs.snowflake.com/en/developer-guide/snowpark/scala/working-with-dataframes@@ -147,5 +147,5 @@    // Create a DataFrame that contains the id, name, and serial_number-  // columns in te "sample_product_data" table.+  // columns in the "sample_product_data" table.   val df = session.table("sample_product_data").select(col("id"), col("name"), col("serial_number"))   df.show()@@ -299,5 +299,5 @@  Note that when specifying the name of a <span className="emph">column</span>, you don't need to use double quotes around the name. The Snowpark library-automatically encloses the column name in double quotes for you if the name does not comply with the identifier requirements:.+automatically encloses the column name in double quotes for you if the name does not comply with the identifier requirements:  ```scala@@ -1306,5 +1306,5 @@ ### Merging Rows into a Table -To insert, update, and deletes rows in one table based on values in a second table or a subquery (the equivalent of the+To insert, update, and delete rows in one table based on values in a second table or a subquery (the equivalent of the [](/sql-reference/sql/merge) command in SQL), do the following: @@ -1318,5 +1318,5 @@    - To specify the update or deletion that should be performed on matching rows, call the `whenMatched` method. -     If you need to specify an additional condition whe rows should be updated or deleted, you can pass in a column expression for+     If you need to specify an additional condition when rows should be updated or deleted, you can pass in a column expression for      that condition.  

--- ahttps://docs.snowflake.com/en/release-notes/bcr-bundles/un-bundled/unbundled-behavior-changes+++ bhttps://docs.snowflake.com/en/release-notes/bcr-bundles/un-bundled/unbundled-behavior-changes@@ -208,4 +208,10 @@     <tr>       <td>**August 20, 2026**</td>+      <td>Azure access: Snowflake Network Identifiers for rules that filter based on subnet IDs</td>+      <td>[](/release-notes/bcr-bundles/un-bundled/bcr-2391)</td>+      <td>Customers who allowlist Azure Storage or Azure Key Vault by subnet ID must also allowlist Snowflake Network Identifiers, preferably through Azure Network Security Perimeter (NSP). Key Vault requires NSP; Storage can use NSP or the storage firewall during migration.</td>+    </tr>+    <tr>+      <td>**August 20, 2026**</td>       <td>Cortex Agents — Managed MCP Server</td>       <td>[](/release-notes/bcr-bundles/un-bundled/bcr-2405)</td> 

--- ahttps://docs.snowflake.com/en/developer-guide/odbc/odbc+++ bhttps://docs.snowflake.com/en/developer-guide/odbc/odbc@@ -1,3 +1,7 @@ # ODBC Driver++Snowflake ODBC Driver 4.x, built on the Universal Core, is in public preview. Installing it replaces the 3.x driver on that machine, so validate it on a separate host. This page documents 3.x, which remains the default and the version Snowflake recommends for production.++See [](/developer-guide/odbc/odbc-universal-core).  - [](/release-notes/client-change-log-odbc)@@ -5,4 +9,5 @@  - [](/release-notes/clients-drivers/odbc)+- [](/developer-guide/odbc/odbc-universal-core)  Snowflake provides a driver for connecting to Snowflake using ODBC-based client applications.@@ -26,2 +31,3 @@ - [](/developer-guide/odbc/odbc-using) - [](/developer-guide/odbc/odbc-diagnostic-service)+- [](/developer-guide/odbc/odbc-universal-core) 

--- ahttps://docs.snowflake.com/en/sql-reference/functions/get_ddl+++ bhttps://docs.snowflake.com/en/sql-reference/functions/get_ddl@@ -31,8 +31,9 @@ - Online feature tables (see [](/sql-reference/sql/create-online-feature-table)) - Pipes (see [](/sql-reference/sql/create-pipe))-- Policies (see [](/sql-reference/sql/create-aggregation-policy) , [](/sql-reference/sql/create-authentication-policy) , [](/sql-reference/sql/create-join-policy) ,+- Policies (see [](/sql-reference/sql/create-aggregation-policy) , [](/sql-reference/sql/create-authentication-policy) , [](/sql-reference/sql/create-data-movement-policy) , [](/sql-reference/sql/create-join-policy) ,   [](/sql-reference/sql/create-masking-policy) , [](/sql-reference/sql/create-password-policy) , [](/sql-reference/sql/create-privacy-policy) ,   [](/sql-reference/sql/create-projection-policy) , [](/sql-reference/sql/create-row-access-policy) , [](/sql-reference/sql/create-session-policy),   [](/sql-reference/sql/create-storage-lifecycle-policy))+- Data movement rules (see [](/sql-reference/sql/create-data-movement-rule)) - Replication groups (see [](/sql-reference/sql/create-replication-group)) - Schemas (see [](/sql-reference/sql/create-schema))@@ -69,4 +70,5 @@ - CONTACT - DATABASE+- DATA MOVEMENT RULE - DYNAMIC_TABLE - EVENT_TABLE@@ -77,5 +79,5 @@ - INTEGRATION (storage) - PIPE-- POLICY (aggregation, authentication, join, masking, password, projection, row access, session, and storage lifecycle policies)+- POLICY (aggregation, authentication, data movement, join, masking, password, projection, row access, session, and storage lifecycle policies) - PROCEDURE (for stored procedures) - REPLICATION_GROUP 

--- ahttps://docs.snowflake.com/en/developer-guide/snowpark/scala/calling-functions+++ bhttps://docs.snowflake.com/en/developer-guide/snowpark/scala/calling-functions@@ -107,5 +107,5 @@    values. -`table?Function` returns a DataFrame that contains the output of the table function.+The `session.tableFunction` method returns a DataFrame that contains the output of the table function.  For example, suppose that you executed the following command to create a SQL UDTF:@@ -207,8 +207,6 @@   ) +// Execute the procedure on the server by passing the procedure's name. session.storedProcedure(name, 1).show()--// Execute the procedure on the server by passing the procedure's name.-session.storedProcedure(incrementProc, 1).show();  // Execute the procedure on the server by passing a variable 

--- ahttps://docs.snowflake.com/en/developer-guide/snowpark/python/snowflake-extension-checkpoints+++ bhttps://docs.snowflake.com/en/developer-guide/snowpark/python/snowflake-extension-checkpoints@@ -1,13 +1,13 @@ <a id="label-snowpark-python-checkpoints-extension"></a> -# Install the Snowflake extension for Snowpark Checkpoints+# Install the Snowflake Extension for Snowpark Checkpoints  ## Prerequisites -Before you use the Checkpoints feature in the Snowflake extension, you must [install the Snowpark Checkpoints library](/developer-guide/snowpark/python/checkpoints-installation).+Before you use the Checkpoints feature in the Snowflake Extension, you must [install the Snowpark Checkpoints library](/developer-guide/snowpark/python/checkpoints-installation).  ## Install the extension -- To use Snowpark Checkpoints from the VS Code Snowflake extension, install the extension from the Visual Studio Code Marketplace:+- To use Snowpark Checkpoints from the Snowflake Extension for Visual Studio Code, install the extension from the Visual Studio Code Marketplace:  ![Checkpoints VSCode extension](/static/images/checkpoints-vscode.png) 

--- ahttps://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-agents-evaluations+++ bhttps://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-agents-evaluations@@ -696,5 +696,5 @@ #### Metrics configuration example -The following example defines a metrics configuration that enables answer correctness and logical consistency checks, and also defines a custom `relevance` metric which returns a score between 1-10 based on how ground truth compares against agent output:+The following example defines a metrics configuration that enables answer correctness and logical consistency checks, and also defines a custom `relevance` metric which returns a score between 1 and 10 based on how ground truth compares against agent output:  ```yaml@@ -796,5 +796,5 @@ ``` -Upload your configuration to a created stage through %sf-web-interface% by navigating to In the navigation menu, select **Ingestion** %raa% **Add Data** and selecting **Load files into a Stage**. You can also use the SQL [](/sql-reference/sql/put) command to upload a local YAML file. The following example demonstrates copying the local file `/Users/dev/evaluation_config.yaml` to the stage `evals_db.evals_schema.evaluation_config`:+Upload your configuration to a created stage through %sf-web-interface% by selecting **Ingestion** %raa% **Add Data**, then **Load files into a Stage**. You can also use the SQL [](/sql-reference/sql/put) command to upload a local YAML file. The following example demonstrates copying the local file `/Users/dev/evaluation_config.yaml` to the stage `evals_db.evals_schema.evaluation_config`:  ```sql@@ -971,5 +971,5 @@ - **Skills**: Evaluations don't currently support skills. An evaluation of an agent that uses skills fails to run. -## Cost Considerations+## Cost considerations  Agent Evaluations run a Cortex Agent to create output for evaluation, and LLM judges to compute the evaluation metrics. You're charged for each run of the agent against a ground truth query. The evaluation's LLM judges are run by the [](/sql-reference/functions/ai_complete) function, and you incur charges based on the model Snowflake selects for judging. Additionally, you're charged for the following: 

--- ahttps://docs.snowflake.com/en/developer-guide/snowpark/java/creating-udfs+++ bhttps://docs.snowflake.com/en/developer-guide/snowpark/java/creating-udfs@@ -291,5 +291,5 @@ UserDefinedFunction langUdf =   Functions.udf(-    (String s) -> Option(detector.detect(s)).getOrElse("UNKNOWN"),+    (String s) -> java.util.Optional.ofNullable(detector.detect(s)).orElse("UNKNOWN"),     DataTypes.StringType,     DataTypes.StringType);@@ -298,5 +298,5 @@ // detected by the UDF. DataFrame dfEmailsWithLangCol =-    dfEmails.withColumn("lang", langUdf(Functions.col("text_data")));+    dfEmails.withColumn("lang", langUdf.apply(Functions.col("text_data"))); ``` @@ -451,5 +451,4 @@       );     dummy.select(udfOnceInit.apply(dummy.col("c"))).show();-    UserDefinedFunction udfOnceInit = udf((i: Int) => sC.randomInt);   } } 

--- ahttps://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-agents-monitor+++ bhttps://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-agents-monitor@@ -72,6 +72,6 @@ 1. Sign in to %sf-web-interface-link%. 2. In the navigation menu, select **AI & ML** %raa% **Agents**.-3. Select the Agent whose logs you wish to view.-4. Navigate to the **Monitoring** pane of the Agent view.+3. Select the agent whose logs you want to view.+4. Navigate to the **Monitoring** pane of the agent view.  The **Monitoring** pane lists each conversation thread with its first input, user feedback, thread length, the user who ran it, and when it was last updated. The **Version** column shows the agent configuration version that handled each thread, so you can tell which version of the agent produced a given conversation. 

--- ahttps://docs.snowflake.com/en/developer-guide/python-connector/python-connector+++ bhttps://docs.snowflake.com/en/developer-guide/python-connector/python-connector@@ -1,3 +1,7 @@ # Snowflake Connector for Python++Snowflake Connector for Python 5.x, built on the Universal Core, is in public preview. It keeps the PEP 249 interface, the `snowflake.connector` module name, and the `snowflake-connector-python` package name. This page documents 4.x, which remains the default and the version Snowflake recommends for production.++See [](/developer-guide/python-connector/python-connector-universal-core).  - [](/release-notes/requirements) 

--- ahttps://docs.snowflake.com/en/release-notes/requirements+++ bhttps://docs.snowflake.com/en/release-notes/requirements@@ -1,3 +1,7 @@ # Client versions & support policy++Snowflake ODBC Driver 4.x and Snowflake Connector for Python 5.x are in public preview. Preview versions are not listed on this page and are not covered by the support policy described here. The recommended and minimum supported versions below refer to the generally available drivers.++See [](/developer-guide/universal-core/universal-core).  - [](/guides-overview-connecting) 

--- ahttps://docs.snowflake.com/en/user-guide/tables-iceberg-access-using-external-query-engine-snowflake-horizon+++ bhttps://docs.snowflake.com/en/user-guide/tables-iceberg-access-using-external-query-engine-snowflake-horizon@@ -13,4 +13,8 @@ You can use Horizon Catalog, which is available in all your existing Snowflake accounts, to read and write to Snowflake-managed Iceberg tables with external query engines.++You can also access externally managed Iceberg tables in a [catalog-linked database](/user-guide/tables-iceberg-catalog-linked-database)+through the same Horizon IRC endpoint. For instructions, see+[](/user-guide/externally-managed-iceberg-tables-access-horizon-irc).  ## Query Iceberg tables 

--- ahttps://docs.snowflake.com/en/developer-guide/odbc/odbc-download+++ bhttps://docs.snowflake.com/en/developer-guide/odbc/odbc-download@@ -1,3 +1,7 @@ # Downloading the ODBC Driver++The downloads on this page are for the 3.x driver.++Snowflake ODBC Driver 4.x, built on the Universal Core, is in public preview and is downloaded separately. See [Installing the preview driver](/developer-guide/odbc/odbc-universal-core#label-odbc-universal-core-install).  - [](/release-notes/client-change-log-odbc) 

--- ahttps://docs.snowflake.com/en/developer-guide/python-connector/python-connector-install+++ bhttps://docs.snowflake.com/en/developer-guide/python-connector/python-connector-install@@ -1,3 +1,7 @@ # Installing the Python Connector++The instructions on this page install the 4.x connector.++Snowflake Connector for Python 5.x, built on the Universal Core, is in public preview. Because both lines use the `snowflake-connector-python` package name, it installs into a separate virtual environment. See [Installing the release candidate](/developer-guide/python-connector/python-connector-universal-core#label-python-universal-core-install).  - [](/release-notes/requirements) 

--- ahttps://docs.snowflake.com/en/developer-guide/drivers+++ bhttps://docs.snowflake.com/en/developer-guide/drivers@@ -1,3 +1,7 @@ # Drivers++New major versions of two Snowflake drivers are in public preview: [Snowflake ODBC Driver 4.x](/developer-guide/odbc/odbc-universal-core) and [Snowflake Connector for Python 5.x](/developer-guide/python-connector/python-connector-universal-core). Both are built on the Universal Core, a shared Rust library used by every new Snowflake driver. The generally available drivers documented in this section remain the default.++See [](/developer-guide/universal-core/universal-core).  - [Developer Resources](https://developers.snowflake.com/resources/)