Scan vs query performance in AWS - CLI Comparison

Introduction

When you want to find data in a database, you can either look through everything or search smartly. Scanning checks every item, which can be slow. Querying looks only where it needs to, making it faster and cheaper.

When you need to find a few specific items quickly in a large database table.

When you want to avoid reading the whole table to save time and cost.

When you have a known key or index to search by in your data.

When you want to get all items regardless of keys, but understand it will be slower.

When you want to filter data but don't have a key to query on.

Commands

This command scans the entire MusicCollection table and returns up to 5 items. It reads all data, which can be slow for big tables.

Terminal

aws dynamodb scan --table-name MusicCollection --limit 5

Expected Output

{

    "Items": [

        {"Artist": {"S": "No One You Know"}, "SongTitle": {"S": "Call Me Today"}},

        {"Artist": {"S": "Acme Band"}, "SongTitle": {"S": "Happy Day"}},

        {"Artist": {"S": "The Wild Sax Band"}, "SongTitle": {"S": "Take Five"}},

        {"Artist": {"S": "No One You Know"}, "SongTitle": {"S": "I Want It That Way"}},

        {"Artist": {"S": "Acme Band"}, "SongTitle": {"S": "Another One Bites The Dust"}}

    ],

    "Count": 5,

    "ScannedCount": 5

}

→

--table-name - Specifies the DynamoDB table to scan

→

--limit - Limits the number of items returned

This command queries the MusicCollection table for items where the Artist is 'No One You Know'. It reads only matching items, making it faster and cheaper.

Terminal

aws dynamodb query --table-name MusicCollection --key-condition-expression "Artist = :artist" --expression-attribute-values '{":artist":{"S":"No One You Know"}}' --limit 5

Expected Output

{

    "Items": [

        {"Artist": {"S": "No One You Know"}, "SongTitle": {"S": "Call Me Today"}},

        {"Artist": {"S": "No One You Know"}, "SongTitle": {"S": "I Want It That Way"}}

    ],

    "Count": 2,

    "ScannedCount": 2

}

→

--key-condition-expression - Defines the condition to find matching keys

→

--expression-attribute-values - Provides values for the condition placeholders

→

--limit - Limits the number of items returned

Key Concept

If you remember nothing else, remember: querying uses keys to find data fast, scanning reads everything and is slower.

Common Mistakes

Using scan when you can use query with a key

Scan reads the whole table, wasting time and money when a query could find data faster.

Use query with key conditions whenever possible to improve performance.

Trying to query without specifying key conditions

Query requires a key condition; without it, the command fails or returns no data.

Always provide a key condition expression when using query.

Summary

Scan reads every item in the table and is slower and more expensive.

Query uses keys to find matching items quickly and efficiently.

Use query when you know the key; use scan only when you need to read all data.

Practice

1. Which AWS DynamoDB operation is faster when you know the exact primary key of the item you want to retrieve?

easy

A. Update

B. Scan

C. Query

D. Delete

Solution

1. Step 1: Understand Query operation

   Query retrieves items by specifying the primary key, making it fast and efficient.

2. Step 2: Compare with Scan operation

   Scan reads the entire table, which is slower and more costly when you know the key.

3. Final Answer:

   Query -> Option C

4. Quick Check:

   Known key = Query [OK]

Hint: Use Query for known keys, Scan reads all data [OK]

Common Mistakes:

• Thinking Scan is faster for known keys
• Confusing Update or Delete with data retrieval
• Assuming Query reads the whole table

2. Which of the following is the correct syntax to perform a Query operation in AWS DynamoDB using the AWS CLI?

easy

A. aws dynamodb query --table-name MyTable --key-condition-expression 'Id = :id' --expression-attribute-values '{":id":{"S":"123"}}'

B. aws dynamodb scan --table-name MyTable --key-condition-expression 'Id = :id' --expression-attribute-values '{":id":{"S":"123"}}'

C. aws dynamodb get-item --table-name MyTable --key '{"Id":{"S":"123"}}'

D. aws dynamodb update-item --table-name MyTable --key '{"Id":{"S":"123"}}'

Solution

1. Step 1: Identify Query command syntax

   The Query operation uses 'aws dynamodb query' with key-condition-expression and expression-attribute-values.

2. Step 2: Check other options

   Scan reads the entire table (and does not support --key-condition-expression), get-item retrieves a single item by exact key, update-item modifies items, so only the Query syntax is correct.

3. Final Answer:

   aws dynamodb query --table-name MyTable --key-condition-expression 'Id = :id' --expression-attribute-values '{":id":{"S":"123"}}' -> Option A

4. Quick Check:

   Query syntax uses 'query' command [OK]

Hint: Query uses 'aws dynamodb query' command [OK]

Common Mistakes:

• Using Scan command instead of Query
• Confusing GetItem with Query
• Incorrect expression-attribute-values format

3. Given a DynamoDB table with 1000 items, what will be the main difference in performance between these two operations?

1. Query with KeyConditionExpression on partition key
2. Scan without any filter

medium

A. Both read the same number of items and have similar speed

B. Scan reads fewer items and is faster; Query reads all items and is slower

C. Both read all items but Query is slower due to filtering

D. Query reads fewer items and is faster; Scan reads all items and is slower

Solution

1. Step 1: Analyze Query operation

   Query uses the partition key to directly find matching items, reading fewer items and faster.

2. Step 2: Analyze Scan operation

   Scan reads every item in the table, making it slower and more expensive.

3. Final Answer:

   Query reads fewer items and is faster; Scan reads all items and is slower -> Option D

4. Quick Check:

   Query = fewer reads, Scan = full table [OK]

Hint: Query targets keys, Scan reads whole table [OK]

Common Mistakes:

• Assuming Scan is faster than Query
• Thinking Query reads all items
• Ignoring cost differences

4. You wrote this DynamoDB Scan command but it is very slow:

aws dynamodb scan --table-name Users

What is the best way to improve performance if you only want users with UserId = '123'?

medium

A. Replace Scan with Query using KeyConditionExpression on UserId

B. Add a FilterExpression to the Scan command

C. Increase the Read Capacity Units of the table

D. Use UpdateItem to modify the UserId

Solution

1. Step 1: Identify problem with Scan

   Scan reads all items, causing slowness when filtering after reading.

2. Step 2: Use Query for known key

   Query with KeyConditionExpression on UserId reads only matching items, improving speed.

3. Final Answer:

   Replace Scan with Query using KeyConditionExpression on UserId -> Option A

4. Quick Check:

   Known key = Query, not Scan [OK]

Hint: Use Query with key condition, not Scan with filter [OK]

Common Mistakes:

• Adding filter to Scan instead of using Query
• Increasing capacity without optimizing query
• Using UpdateItem unrelated to read performance

5. You have a DynamoDB table with 1 million items. You need to retrieve all items where the attribute 'Status' equals 'Active'. The table's primary key is 'UserId'. Which approach gives the best performance and cost efficiency?

hard

A. Use Scan with FilterExpression 'Status = Active' to get all matching items

B. Create a Global Secondary Index (GSI) on 'Status' and use Query on the GSI

C. Use Query on the primary key 'UserId' with FilterExpression 'Status = Active'

D. Use Scan without any filter and filter results in your application code

Solution

1. Step 1: Understand Scan with filter

   Scan reads all 1 million items, filtering after reading, which is slow and costly.

2. Step 2: Use GSI for efficient querying

   Creating a GSI on 'Status' allows Query to directly find 'Active' items efficiently.

3. Step 3: Evaluate other options

   Query on primary key can't filter by 'Status' efficiently; scanning without filter wastes resources.

4. Final Answer:

   Create a Global Secondary Index (GSI) on 'Status' and use Query on the GSI -> Option B

5. Quick Check:

   GSI + Query = best performance for non-key attribute [OK]

Hint: Use GSI and Query for non-key attribute filtering [OK]

Common Mistakes:

• Using Scan with filter on large tables
• Querying primary key with filter instead of GSI
• Filtering in application after full Scan