Undrop for recovering dropped objects in Snowflake - Commands & Configuration

Introduction

Sometimes you accidentally delete a table or other database object. Snowflake lets you recover these dropped objects quickly using the undrop command, so you don't lose important data.

When you accidentally drop a table but realize you still need the data.

When a colleague drops a view by mistake and you want to restore it.

When you want to recover a schema or database that was dropped recently.

When you want to restore an object to its exact state before deletion.

When you want to avoid recreating complex objects from scratch after accidental deletion.

Commands

This command lists all tables that were dropped recently in the specified schema. It helps you find the exact name of the table you want to recover.

Terminal

SHOW DROPPED TABLES IN SCHEMA my_database.public;

Expected Output

name           | dropped_on          | dropped_by

my_table       | 2024-06-01 10:15:00 | [e-mail]

This command recovers the dropped table named 'my_table' in the specified schema. It restores the table with all its data and structure.

Terminal

UNDROP TABLE my_database.public.my_table;

Expected Output

Statement executed successfully.

This command verifies that the table has been restored and is available again in the schema.

Terminal

SHOW TABLES IN SCHEMA my_database.public;

Expected Output

name           | created_on          | rows

my_table       | 2024-06-01 10:15:00 | 1000

Key Concept

If you remember nothing else, remember: Snowflake keeps dropped objects for a time so you can easily undrop and recover them.

Common Mistakes

Trying to undrop an object without specifying the full name including database and schema.

Snowflake needs the full path to find the correct dropped object to restore.

Always use the full object name like database.schema.object when running undrop.

Waiting too long to undrop, after the retention period expires.

Dropped objects are only kept for a limited time before permanent deletion.

Run undrop as soon as possible after accidental drop to ensure recovery.

Summary

Use SHOW DROPPED TABLES to find recently dropped tables in a schema.

Run UNDROP TABLE with full object name to recover the dropped table.

Verify recovery by listing tables again with SHOW TABLES.

Practice

1. What is the primary purpose of the UNDROP command in Snowflake?

easy

A. To recover a dropped table, view, or schema within the retention period

B. To permanently delete a table or schema from Snowflake

C. To create a new table with the same structure as an existing one

D. To rename an existing table or schema

Solution

1. Step 1: Understand the function of UNDROP

   The UNDROP command is designed to restore objects that were accidentally dropped, such as tables, views, or schemas.

2. Step 2: Identify the correct use case

   Since UNDROP recovers dropped objects within the retention period, it is not used for deletion, creation, or renaming.

3. Final Answer:

   To recover a dropped table, view, or schema within the retention period -> Option A

4. Quick Check:

   UNDROP = Recover dropped objects [OK]

Hint: UNDROP means undo drop to restore deleted objects [OK]

Common Mistakes:

• Confusing UNDROP with DELETE or DROP commands
• Thinking UNDROP creates new objects
• Assuming UNDROP renames objects

2. Which of the following is the correct syntax to undrop a table named sales_data in Snowflake?

easy

A. UNDROP TABLE sales_data;

B. RECOVER TABLE sales_data;

C. UNDROP TABLE IF EXISTS sales_data;

D. RESTORE TABLE sales_data;

Solution

1. Step 1: Recall the correct UNDROP syntax

   The correct syntax to recover a dropped table is UNDROP TABLE <table_name>;

2. Step 2: Eliminate incorrect options

   Adding IF EXISTS is invalid with UNDROP. RECOVER and RESTORE are not supported commands for recovering dropped tables in Snowflake.

3. Final Answer:

   UNDROP TABLE sales_data; -> Option A

4. Quick Check:

   UNDROP TABLE = correct syntax [OK]

Hint: Use UNDROP TABLE followed by the table name [OK]

Common Mistakes:

• Using RECOVER or RESTORE instead of UNDROP
• Adding IF EXISTS which is invalid with UNDROP
• Missing the TABLE keyword

3. Consider the following sequence of commands in Snowflake:

DROP TABLE IF EXISTS customers;
UNDROP TABLE customers;

What will be the result of the UNDROP TABLE customers; command assuming the drop happened within the retention period?

medium

A. An error will occur because the table does not exist

B. The command will be ignored silently

C. The table will be restored but without any data

D. The table 'customers' will be restored with its data intact

Solution

1. Step 1: Understand the effect of DROP and UNDROP

   The DROP TABLE command deletes the table but keeps it recoverable within the retention period. UNDROP restores the dropped table with its data.

2. Step 2: Confirm retention period and command behavior

   Since the drop is recent and within retention, UNDROP will successfully restore the table and its data.

3. Final Answer:

   The table 'customers' will be restored with its data intact -> Option D

4. Quick Check:

   UNDROP restores dropped table and data [OK]

Hint: UNDROP restores table and data if within retention [OK]

Common Mistakes:

• Assuming UNDROP only restores structure, not data
• Thinking UNDROP fails if table was dropped
• Believing UNDROP silently ignores missing tables

4. You tried to undrop a dropped schema using UNDROP SCHEMA sales_schema; but received an error. Which of the following is the most likely cause?

medium

A. You used the wrong command; UNDROP does not support schemas

B. The schema was dropped outside the retention period

C. The schema name is case-sensitive and was typed incorrectly

D. You need to run UNDROP with the database name specified

Solution

1. Step 1: Check retention period requirement

   UNDROP only works if the object was dropped within the retention period. If dropped too long ago, recovery fails.

2. Step 2: Verify command support and syntax

   UNDROP supports schemas, and schema names are case-insensitive by default. Database name is not required if context is set.

3. Final Answer:

   The schema was dropped outside the retention period -> Option B

4. Quick Check:

   Retention period exceeded causes UNDROP error [OK]

Hint: Check retention period before undropping schemas [OK]

Common Mistakes:

• Assuming UNDROP can't recover schemas
• Ignoring retention period limits
• Mistyping schema name without checking case sensitivity

5. You accidentally dropped a table orders in your production database. You want to recover it using UNDROP, but you also want to ensure no data loss and minimal downtime. Which of the following steps is the best approach?

hard

A. Wait for the retention period to expire and then restore from a full database backup

B. Recreate the table manually and import data from backups

C. Immediately run UNDROP TABLE orders; within the retention period and verify data integrity

D. Use UNDROP TABLE orders; only after recreating the table structure

Solution

1. Step 1: Use UNDROP immediately after accidental drop

   UNDROP recovers the dropped table with data if done within the retention period, minimizing downtime and data loss.

2. Step 2: Avoid manual recreation or waiting unnecessarily

   Manual recreation or waiting for backup restore increases downtime and risks data loss.

3. Final Answer:

   Immediately run UNDROP TABLE orders; within the retention period and verify data integrity -> Option C

4. Quick Check:

   UNDROP quickly recovers dropped tables with data [OK]

Hint: Run UNDROP immediately after drop to avoid data loss [OK]

Common Mistakes:

• Delaying UNDROP beyond retention period
• Recreating table manually before undropping
• Ignoring data verification after recovery