Warehouse selection strategies in Snowflake - Commands & Configuration

Introduction

When you run queries in Snowflake, you need to choose a warehouse. Warehouses are like engines that process your data. Choosing the right warehouse helps your queries run faster and saves money.

When you want to run a quick report without waiting too long.

When you have many users running queries at the same time.

When you want to save money by using smaller warehouses for light tasks.

When you need to run heavy data processing jobs quickly.

When you want to balance speed and cost by switching warehouses based on workload.

Commands

This command sets the current warehouse to COMPUTE_WH_SMALL. It tells Snowflake to use this warehouse for running your queries.

Terminal

USE WAREHOUSE COMPUTE_WH_SMALL;

Expected Output

No output (command runs silently)

This command shows which warehouse is currently active. It helps you confirm that your selection worked.

Terminal

SELECT CURRENT_WAREHOUSE();

Expected Output

CURRENT_WAREHOUSE()

COMPUTE_WH_SMALL

This command changes the warehouse for the current session to COMPUTE_WH_LARGE. Use this when you need more power for big queries.

Terminal

ALTER SESSION SET WAREHOUSE = COMPUTE_WH_LARGE;

Expected Output

No output (command runs silently)

This command lists all warehouses in your Snowflake account. It helps you see available options and their sizes.

Terminal

SHOW WAREHOUSES;

Expected Output

name              | state   | size    | type     | auto_suspend | auto_resume

COMPUTE_WH_SMALL  | started | X-Small | standard | 300          | true

COMPUTE_WH_LARGE  | started | Large   | standard | 300          | true

Key Concept

If you remember nothing else from this pattern, remember: choosing the right warehouse size balances query speed and cost.

Common Mistakes

Not setting a warehouse before running queries.

Queries will fail because Snowflake does not know which compute resource to use.

Always run USE WAREHOUSE command or set warehouse in your session before querying.

Using a large warehouse for small queries.

This wastes money because large warehouses cost more to run.

Use smaller warehouses for light workloads to save cost.

Forgetting to check which warehouse is active.

You might run queries on the wrong warehouse, causing unexpected cost or slow performance.

Run SELECT CURRENT_WAREHOUSE() to confirm your warehouse before running queries.

Summary

Use USE WAREHOUSE command to select the warehouse for your queries.

Check the active warehouse with SELECT CURRENT_WAREHOUSE().

Switch warehouses with ALTER SESSION SET WAREHOUSE for different workloads.

List all warehouses with SHOW WAREHOUSES to choose the right one.

Practice

1. What does the USE WAREHOUSE command do in Snowflake?

easy

A. It lists all available warehouses.

B. It switches the current warehouse to run queries.

C. It deletes an existing warehouse.

D. It creates a new warehouse.

Solution

1. Step 1: Understand the purpose of USE WAREHOUSE

   This command changes which warehouse is active for running queries.

2. Step 2: Compare with other commands

   Creating, deleting, or listing warehouses use different commands, not USE WAREHOUSE.

3. Final Answer:

   It switches the current warehouse to run queries. -> Option B

4. Quick Check:

   USE WAREHOUSE changes active warehouse [OK]

Hint: Remember: USE means switch active warehouse [OK]

Common Mistakes:

• Confusing USE with CREATE or DROP commands
• Thinking it lists warehouses
• Assuming it modifies warehouse size

2. Which of the following is the correct syntax to switch to a warehouse named COMPUTE_WH?

easy

A. USE WAREHOUSE COMPUTE_WH;

B. SWITCH WAREHOUSE COMPUTE_WH;

C. SELECT WAREHOUSE COMPUTE_WH;

D. CHANGE WAREHOUSE TO COMPUTE_WH;

Solution

1. Step 1: Recall the correct command syntax

   The command to switch warehouses is USE WAREHOUSE <name>;.

2. Step 2: Check each option

   Only USE WAREHOUSE COMPUTE_WH; matches the correct syntax exactly.

3. Final Answer:

   USE WAREHOUSE COMPUTE_WH; -> Option A

4. Quick Check:

   Correct syntax = USE WAREHOUSE [name]; [OK]

Hint: USE WAREHOUSE plus name switches warehouse [OK]

Common Mistakes:

• Using SWITCH or CHANGE instead of USE
• Adding extra keywords like TO or SELECT
• Missing semicolon at end

3. Given the following commands run in order:

USE WAREHOUSE SMALL_WH;
SELECT CURRENT_WAREHOUSE();
USE WAREHOUSE LARGE_WH;
SELECT CURRENT_WAREHOUSE();

What will be the output of the two CURRENT_WAREHOUSE() calls?

medium

A. SMALL_WH then LARGE_WH

B. LARGE_WH then SMALL_WH

C. SMALL_WH twice

D. LARGE_WH twice

Solution

1. Step 1: Analyze the first USE WAREHOUSE command

   The first command sets the active warehouse to SMALL_WH, so the first CURRENT_WAREHOUSE() returns SMALL_WH.

2. Step 2: Analyze the second USE WAREHOUSE command

   The second command switches the active warehouse to LARGE_WH, so the second CURRENT_WAREHOUSE() returns LARGE_WH.

3. Final Answer:

   SMALL_WH then LARGE_WH -> Option A

4. Quick Check:

   Warehouse switches reflect in CURRENT_WAREHOUSE() [OK]

Hint: Each USE WAREHOUSE changes current warehouse immediately [OK]

Common Mistakes:

• Assuming warehouse does not change after second USE
• Mixing order of outputs
• Thinking CURRENT_WAREHOUSE() returns session name

4. You run the command USE WAREHOUSE BIG_WH but get an error saying the warehouse does not exist. What is the most likely cause?

medium

A. The command requires a semicolon at the end.

B. You need to restart your Snowflake session.

C. You must create a database before switching warehouses.

D. The warehouse name is misspelled or does not exist.

Solution

1. Step 1: Understand the error message

   The error says the warehouse does not exist, so the name is likely wrong or missing.

2. Step 2: Evaluate other options

   Restarting session or creating a database is unrelated; semicolon missing usually causes syntax error, not 'does not exist'.

3. Final Answer:

   The warehouse name is misspelled or does not exist. -> Option D

4. Quick Check:

   Nonexistent warehouse name causes error [OK]

Hint: Check warehouse name spelling if 'does not exist' error appears [OK]

Common Mistakes:

• Restarting session unnecessarily
• Confusing warehouse with database
• Ignoring error message details

5. You have two warehouses: SMALL_WH with 1 cluster and LARGE_WH with 4 clusters. You want to run a heavy query quickly but also save costs on light queries. Which strategy is best?

hard

A. Always use LARGE_WH for all queries to ensure speed.

B. Use SMALL_WH only and increase its size when needed.

C. Use SMALL_WH for light queries and switch to LARGE_WH for heavy queries.

D. Create a new warehouse for each query type.

Solution

1. Step 1: Understand warehouse sizes and costs

   SMALL_WH is cheaper and good for light work; LARGE_WH is bigger and faster but costs more.

2. Step 2: Choose strategy balancing speed and cost

   Switching warehouses based on query load saves cost and improves performance.

3. Final Answer:

   Use SMALL_WH for light queries and switch to LARGE_WH for heavy queries. -> Option C

4. Quick Check:

   Match warehouse size to query load for cost and speed [OK]

Hint: Match warehouse size to query weight, switch as needed [OK]

Common Mistakes:

• Using large warehouse always wastes cost
• Trying to resize warehouses dynamically
• Creating many warehouses unnecessarily