Restores the most recent version of a dropped table.
CREATE TABLE , ALTER TABLE , DROP TABLE , SHOW TABLES , DESCRIBE TABLE
name
Specifies the identifier for the table to restore. If the identifier contains spaces or special characters, the entire string must be enclosed in double quotes. Identifiers enclosed in double quotes are also case-sensitive.
Tables can only be restored to the database and schema that contained the table at the time of deletion. For example, if you create and drop table t1 in schema s1, then change the current schema to s2 and attempt to restore table t1 by ID (or qualified name, s1.t1), table t1 is restored in schema s1 rather than in the current schema, s2.
If a table with the same name already exists, an error is returned.
If you have multiple dropped tables with the same name, you can use the IDENTIFIER keyword with the system-generated identifier (from the TABLES view) to specify which table to restore. The name of the restored table remains the same. See Examples.
Note
You can only use the system-generated identifier with the IDENTIFIER() keyword when executing the UNDROP command for notebooks, tables, schemas, and databases.
UNDROP relies on the Snowflake Time Travel feature. An object can be restored only if the object was deleted within the Data retention period. The default value is 24 hours.
You cannot undrop a hybrid table.
Restore the most recent version of a dropped table (this example builds on the examples provided for DROP TABLE):
+---------------------------------+ | status | |---------------------------------| | Table T2 successfully restored. | +---------------------------------+UNDROP table using the table ID¶
Restore a dropped table by ID using IDENTIFIER(). You can find the table ID of the specific table to undrop using the table_id column in the TABLES view. For example, if you have multiple dropped tables named my_table, and you want to restore the second-to-last dropped table my_table, follow these steps:
Find the table ID of the dropped table in the Account Usage TABLES view:
SELECT table_id, table_name, table_schema, table_catalog, created, deleted, comment FROM SNOWFLAKE.ACCOUNT_USAGE.TABLES WHERE table_catalog = 'DB1' AND table_schema = 'S1' AND table_name = 'MY_TABLE' AND deleted IS NOT NULL ORDER BY deleted;
Copy
+----------+------------+--------------+---------------+-------------------------------+-------------------------------+---------+ | TABLE_ID | TABLE_NAME | TABLE_SCHEMA | TABLE_CATALOG | CREATED | DELETED | COMMENT | |----------+------------+--------------+---------------+-------------------------------+-------------------------------+---------| | 408578 | MY_TABLE | S1 | DB1 | 2024-07-01 15:39:07.565 -0700 | 2024-07-01 15:40:28.161 -0700 | NULL | +----------+------------+--------------+---------------+-------------------------------+-------------------------------+---------+ | 408607 | MY_TABLE | S1 | DB1 | 2024-07-01 17:43:07.565 -0700 | 2024-07-01 17:44:28.161 -0700 | NULL | +----------+------------+--------------+---------------+-------------------------------+-------------------------------+---------+
Undrop my_table by table ID. To restore the second-to-last deleted table, use table ID 408578 from the output of the previous statement. After you execute the following statement, the table is restored with its original name, my_table:
UNDROP TABLE IDENTIFIER(408578);
Copy
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4