Error code reference for dynamic tables¶

When a dynamic table refresh fails, Snowflake records an error code and message in the refresh history. Use this page to look up the error, understand the cause, and resolve it.

To find the error code for a failed refresh, run:

SELECT name,
       state,
       state_code,
       state_message,
       query_id
FROM TABLE(INFORMATION_SCHEMA.DYNAMIC_TABLE_REFRESH_HISTORY(
  NAME => 'my_database.my_schema.dt_orders'
))
WHERE state = 'FAILED'
ORDER BY data_timestamp DESC
LIMIT 10;

+----------------------------+---------+------------+--------------------------------------------------+--------------------------------------+
| NAME                       | STATE   | STATE_CODE | STATE_MESSAGE                                    | QUERY_ID                             |
+----------------------------+---------+------------+--------------------------------------------------+--------------------------------------+
| MY_DATABASE...DT_ORDERS    | FAILED  | 091930     | Change tracking is not enabled or has been ...   | 01b23c4d-0501-a2b3-0000-00012345abcd |
+----------------------------+---------+------------+--------------------------------------------------+--------------------------------------+

The state_code column contains the numeric error code. The state_message column contains the full error message. Use the query_id to inspect the failed statement in the query history.

The messages listed below are representative. Your actual error message may include additional details such as object names, timestamps, or incident IDs.

Errors fall into three categories:

• Configuration errors are preventable. They result from missing prerequisites such as change tracking, warehouses, or privileges.
• Operational errors are recoverable. They result from upstream changes, timeouts, or schema evolution that you can fix without Snowflake support.
• Internal errors are not user-actionable. Contact Snowflake Support and include the query_id from the refresh history.

Configuration errors (preventable)¶

These errors indicate that a required setting, object, or privilege is missing. Fix the configuration to resolve the error.

091930¶

Message: Change tracking is not enabled or has been missing for the time range requested.

Cause: A base table referenced by the dynamic table does not have change tracking enabled, or change tracking history was lost.

Resolution: Enable change tracking on the affected base table with ALTER TABLE ... SET CHANGE_TRACKING = TRUE. If the error persists, recreate the dynamic table. For step-by-step diagnosis, see Troubleshoot dynamic table refresh issues.

002725¶

Message: Dynamic table could not be refreshed because warehouse is missing.

Cause: The warehouse assigned to the dynamic table was dropped, renamed, or the owner role lost USAGE privilege on it.

Resolution: Verify that the warehouse exists with SHOW WAREHOUSES LIKE '<name>' and check privileges with SHOW GRANTS ON WAREHOUSE <name>.

If the warehouse was dropped, reassign a warehouse with ALTER DYNAMIC TABLE <name> SET WAREHOUSE = <warehouse_name>.

002798¶

Message: Dynamic table contains a data dependent '<function_name>' function.

Cause: A masking or row-access policy on a base table passes a column value to a context function such as IS_ROLE_IN_SESSION(column_name), and the dynamic table uses REFRESH_MODE = INCREMENTAL. Policies with constant arguments are compatible with incremental refresh and do not trigger this error.

Resolution: Use REFRESH_MODE = AUTO to let Snowflake resolve to FULL refresh gracefully, or rewrite the policy to use constant arguments instead of column values. For more information, see Dynamic table refresh modes.

002037¶

Message: Object does not exist or not authorized.

Cause: A table, view, or function referenced in the dynamic table’s definition does not exist, or the dynamic table’s owner role does not have the required privileges on it.

Resolution: Verify that all referenced objects exist and the owner role has SELECT privilege. Check the full state_message in the refresh history for the name of the missing object, then verify the object and its grants:

SHOW GRANTS TO ROLE my_dt_owner_role;

For more information, see Dynamic table access control.

002744¶

Message: Dynamic table must use full refresh because a security policy on a base table is not compatible with incremental refresh.

Cause: A row access policy or masking policy on a base table uses constructs that incremental refresh can’t track, forcing the dynamic table to use full refresh.

Resolution: Recreate the dynamic table with REFRESH_MODE = FULL:

CREATE OR REPLACE DYNAMIC TABLE my_database.my_schema.dt_orders
  ...
  REFRESH_MODE = FULL
AS
  SELECT ...;

Alternatively, remove or modify the security policy on the source table so it no longer uses incompatible constructs. REFRESH_MODE can’t be changed after creation: to change it, recreate the dynamic table with CREATE OR REPLACE DYNAMIC TABLE. For more information, see Dynamic table refresh modes.

002756¶

Message: Insufficient privileges to operate on dynamic table. OPERATE privilege is required for manual refresh.

Cause: The role running ALTER DYNAMIC TABLE ... REFRESH does not have the OPERATE privilege on the dynamic table. Only the owner role or a role with an explicit OPERATE grant can trigger a manual refresh.

Resolution: Grant the OPERATE privilege to the role that needs to run manual refreshes:

GRANT OPERATE ON DYNAMIC TABLE my_database.my_schema.dt_orders
  TO ROLE pipeline_admin_role;

OWNERSHIP includes all capabilities of OPERATE. For more information, see Dynamic table access control.

002757¶

Message: Insufficient privileges to operate on dynamic table. OPERATE privilege is required for initial refresh.

Cause: The role that created the dynamic table does not have the OPERATE privilege on one or more upstream dynamic tables in the pipeline. During INITIALIZE = ON_CREATE, Snowflake checks that the creating role can operate on all upstream dynamic tables that need to be refreshed.

Resolution: Grant OPERATE on each upstream dynamic table to the role that will execute the CREATE DYNAMIC TABLE statement. After creation, background refreshes use the owner role and do not require OPERATE on upstream tables:

GRANT OPERATE ON DYNAMIC TABLE my_database.my_schema.dt_upstream
  TO ROLE transform_role;

Alternatively, create the dynamic table with INITIALIZE = ON_SCHEDULE to skip the immediate initial refresh. The background refresh process uses the owner role, which implicitly has OPERATE. For more information, see Dynamic table access control.

002773¶

Message: Owner role of the dynamic table does not exist.

Cause: The role that owns the dynamic table was dropped or no longer exists. Snowflake uses the owner role to run background refreshes, so when the role is missing, no refresh can execute.

Resolution: Transfer ownership to an existing role:

GRANT OWNERSHIP ON DYNAMIC TABLE my_database.my_schema.dt_orders TO ROLE transform_role COPY CURRENT GRANTS;

If the role was dropped accidentally, recreate it and re-grant the required privileges. Verify the current owner with:

SHOW DYNAMIC TABLES LIKE 'dt_orders' IN SCHEMA my_database.my_schema;

The owner column shows which role currently owns the dynamic table. For more information, see Dynamic table access control.

Operational errors (recoverable)¶

These errors result from changes in your environment, such as dropped objects or schema evolution. They require you to take action, but do not need Snowflake support.

091912¶

Message: Query contains a non-deterministic function that change tracking does not support.

Cause: The dynamic table’s definition uses a non-deterministic function in a position where incremental refresh can’t track changes. Timestamp functions (CURRENT_TIMESTAMP, CURRENT_DATE, CURRENT_TIME) are supported in WHERE, HAVING, and QUALIFY clauses. This error occurs when a non-deterministic function like RANDOM() or a session-context function (CURRENT_USER, CURRENT_ROLE) appears in the SELECT list or another unsupported position.

Resolution: Move the non-deterministic function to a WHERE, HAVING, or QUALIFY clause (supported for timestamp functions), remove it from the definition, or recreate the dynamic table with REFRESH_MODE = FULL. For example, CURRENT_TIMESTAMP() in the SELECT list forces full refresh:

CREATE OR REPLACE DYNAMIC TABLE my_database.my_schema.dt_orders
  ...
  REFRESH_MODE = FULL
AS
  SELECT order_id, customer_id, CURRENT_TIMESTAMP() AS refreshed_at
  FROM my_database.my_schema.raw_orders;

If you need a timestamp that records when each row was last refreshed, use METADATA$ROW_LAST_COMMIT_TIME instead of CURRENT_TIMESTAMP(). This built-in column provides the actual commit time of each row and is compatible with incremental refresh.

For more information, see Supported queries for dynamic tables.

002741¶

Message: Dynamic table is not initialized.

Cause: The dynamic table’s initial refresh has not completed. This can happen when the initial refresh fails (for example, due to a privilege issue or query error) or when the dynamic table was recently created and has not been scheduled yet.

Resolution: Check the refresh history for the initial refresh failure:

SELECT state, state_code, state_message
FROM TABLE(INFORMATION_SCHEMA.DYNAMIC_TABLE_REFRESH_HISTORY(
  NAME => 'my_database.my_schema.dt_orders'
))
ORDER BY data_timestamp ASC
LIMIT 5;

Fix the root cause (missing privilege, invalid query, warehouse issue), then trigger a manual refresh:

ALTER DYNAMIC TABLE my_database.my_schema.dt_orders REFRESH;

002742¶

Message: Failed to refresh dynamic table. Error in secure object.

Cause: The dynamic table’s definition references a secure view or secure function that returned an error during refresh. This commonly occurs after the secure object itself is modified in a way that breaks the dynamic table’s query.

Resolution: Verify the secure object’s SQL is valid and that the dynamic table’s owner role has the required privileges:

SHOW GRANTS ON VIEW my_database.my_schema.my_secure_view;
DESC VIEW my_database.my_schema.my_secure_view;

If the secure object’s schema changed, recreate the dynamic table with an updated definition.

002724¶

Message: Refresh was canceled or timed out.

Cause: The refresh query exceeded the statement timeout or was canceled. This occurs when the warehouse is too small for the workload, the definition is too complex, or resource contention caused the query to queue.

Resolution: Consider one or more of these approaches:

1. Increase the warehouse size. See Choose and size warehouses for dynamic tables.
2. Simplify the definition by breaking it into multiple dynamic tables.
3. Increase the STATEMENT_TIMEOUT_IN_SECONDS parameter on the warehouse.

The effective timeout is the most restrictive value set across the account, warehouse, and session levels. Check all three if increasing the warehouse-level setting does not resolve the issue.

000627¶

Message: Statement reached its statement or lock timeout.

Cause: The refresh statement could not acquire a lock on the dynamic table or a referenced object within the lock timeout period. This occurs during periods of high concurrency or when a long-running transaction holds a conflicting lock.

Resolution: Check for long-running transactions that may hold locks:

SHOW LOCKS IN ACCOUNT;
SHOW TRANSACTIONS IN ACCOUNT;

If lock contention is recurring, consider increasing the LOCK_TIMEOUT parameter or staggering refresh schedules across the pipeline.

000603¶

Message: SQL execution canceled.

Cause: The refresh query was canceled before it completed. This occurs when the warehouse auto-suspends while a refresh is in progress, an administrator manually cancels the refresh query, or the system cancels the query due to resource pressure.

Resolution: Check the warehouse AUTO_SUSPEND setting. If the warehouse suspends during refresh windows, increase AUTO_SUSPEND or set it to 0 (never auto-suspend) while refreshes are running:

ALTER WAREHOUSE transform_wh SET AUTO_SUSPEND = 600;

If the query was manually canceled, verify that no automated process or user is canceling active refresh queries.

002753¶

Message: Manual refresh failed because an upstream dynamic table failed to refresh.

Cause: A manual refresh (ALTER DYNAMIC TABLE ... REFRESH) failed because one or more upstream dynamic tables in the pipeline also failed. The upstream failure must be resolved before the downstream refresh can succeed.

Resolution: Identify which upstream dynamic table failed by checking the refresh history for each table in the pipeline:

SELECT name, state, state_code, state_message
FROM TABLE(INFORMATION_SCHEMA.DYNAMIC_TABLE_REFRESH_HISTORY())
WHERE state = 'FAILED'
ORDER BY data_timestamp DESC
LIMIT 20;

Fix the upstream failure first, then retry the manual refresh on the downstream dynamic table:

ALTER DYNAMIC TABLE my_database.my_schema.dt_orders REFRESH;

002766¶

Message: Dynamic table can no longer be refreshed incrementally.

Cause: A structural change in the pipeline, such as adding a LIMIT clause to a view or altering a column in a referenced object, means the dynamic table’s definition can no longer be tracked incrementally. Snowflake detected that the change query is no longer valid.

Resolution: Recreate the dynamic table with CREATE OR REPLACE DYNAMIC TABLE. If the definition no longer supports incremental refresh, set REFRESH_MODE = FULL:

CREATE OR REPLACE DYNAMIC TABLE my_database.my_schema.dt_orders
  ...
  REFRESH_MODE = FULL
AS
  SELECT ...;

To keep incremental refresh, undo the structural change that broke compatibility (for example, remove a LIMIT clause from a referenced view). For details on which query patterns support incremental refresh, see Supported queries for dynamic tables.

002841¶

Message: Upstream table is outside the data retention period. Change tracking data is no longer available.

Cause: An upstream table that the dynamic table depends on fell outside its DATA_RETENTION_TIME_IN_DAYS window. The change tracking data needed for incremental refresh is no longer available, so the dynamic table can’t refresh.

Resolution: The dynamic table must reinitialize. Recreate it with CREATE OR REPLACE DYNAMIC TABLE to reset the change tracking baseline:

CREATE OR REPLACE DYNAMIC TABLE my_database.my_schema.dt_orders
  ...
AS
  SELECT ...;

To prevent this issue in the future:

1. Monitor suspension duration. If a dynamic table is suspended, resume it before the upstream table’s retention period expires.
2. Increase the upstream table’s retention period if needed: ALTER TABLE upstream_table SET DATA_RETENTION_TIME_IN_DAYS = 14.
3. Set up alerts on the DYNAMIC_TABLE_REFRESH_HISTORY function to detect suspensions early.

Internal errors (contact support)¶

These errors indicate an internal issue that you can’t resolve on your own. Contact Snowflake Support and provide the query_id and error details from the refresh history.

002781¶

Message: Internal Dynamic Table refresh error. Invalid DT_ROW_ID.

Cause: An internal data integrity check failed during incremental refresh. Snowflake detected rows in the change delta that do not match existing rows in the dynamic table.

Resolution: This is an internal data integrity error. If it occurs, recreate the dynamic table with CREATE OR REPLACE. If the error persists, contact Snowflake Support with the query_id from the refresh history.

300002, 300010¶

Message: SQL execution internal error: Processing aborted due to error.

Cause: An internal execution error occurred during refresh processing. The error message includes a numeric error code (such as 300002 or 300010) and an incident ID. Other codes in the 300xxx range may also appear.

Resolution: This is not user-actionable. Contact Snowflake Support and include the full error message with the error code and incident ID.

370001¶

Message: ExecRsoProjection: Unresolved connector.

Cause: An internal SQL compiler error occurred while processing the dynamic table’s definition.

Resolution: This is not user-actionable. Contact Snowflake Support and include the query_id. As a temporary workaround, simplifying the definition (for example, breaking a complex JOIN + GROUP BY into two dynamic tables) may avoid the issue.

When to contact Snowflake support¶

Contact Snowflake Support when:

• The error is listed in the Internal errors section above.
• The error message contains “Processing aborted due to error” with an incident number.
• The error message contains “Internal Error” or “incident.”
• Error 002781 (invalid DT_ROW_ID) occurs repeatedly.
• You have followed the resolution steps for a configuration or operational error and the issue persists.

When you open a support case, include:

1. The full error message from the refresh history.
2. The query_id from the state_message or refresh history output.
3. The dynamic table name and database/schema.
4. The output of SHOW DYNAMIC TABLES LIKE '<table_name>'.

What’s next¶

• Troubleshoot dynamic table refresh issues
• Monitor dynamic tables
• Dynamic table refresh modes