Querying GSI in DynamoDB

Introduction

We use Querying GSI to quickly find items in a DynamoDB table using an alternate key. It helps when you want to search by something other than the main key.

You want to find all orders by a specific customer ID instead of the order ID.

You need to get all products in a category quickly.

You want to list all events happening on a certain date.

You want to filter items by a status or tag that is not the main key.

Syntax

DynamoDB

aws dynamodb query \
  --table-name TableName \
  --index-name GSIName \
  --key-condition-expression "GSIPartitionKey = :value" \
  --expression-attribute-values '{":value":{"S":"YourValue"}}'

The --index-name option tells DynamoDB which GSI to use.

The --key-condition-expression filters items by the GSI partition key.

Examples

Finds all orders for customer with ID 'C123' using the CustomerIndex GSI.

DynamoDB

aws dynamodb query \
  --table-name Orders \
  --index-name CustomerIndex \
  --key-condition-expression "CustomerId = :cid" \
  --expression-attribute-values '{":cid":{"S":"C123"}}'

Finds all products in the 'Books' category using the CategoryIndex GSI.

DynamoDB

aws dynamodb query \
  --table-name Products \
  --index-name CategoryIndex \
  --key-condition-expression "Category = :cat" \
  --expression-attribute-values '{":cat":{"S":"Books"}}'

Sample Program

This query finds all music items where the artist is 'The Beatles' using the ArtistIndex GSI.

DynamoDB

aws dynamodb query \
  --table-name Music \
  --index-name ArtistIndex \
  --key-condition-expression "Artist = :artist" \
  --expression-attribute-values '{":artist":{"S":"The Beatles"}}'

Important Notes

Make sure the GSI is created before querying it.

You can only query by the GSI's partition key and optionally its sort key.

Querying a GSI is faster than scanning the whole table.

Summary

Querying a GSI lets you find items using an alternate key.

You specify the GSI name and key condition in the query.

This helps you search data in different ways without scanning the whole table.

Practice

1. What is the main purpose of querying a Global Secondary Index (GSI) in DynamoDB?

easy

A. To delete items from the table

B. To create a backup of the table

C. To update the primary key of items

D. To find items using an alternate key without scanning the whole table

Solution

1. Step 1: Understand what a GSI is used for

   A GSI allows querying data using a different key than the primary key.

2. Step 2: Identify the main benefit of querying a GSI

   Querying a GSI helps find items quickly without scanning the entire table.

3. Final Answer:

   To find items using an alternate key without scanning the whole table -> Option D

4. Quick Check:

   Query GSI = alternate key search [OK]

Hint: GSI lets you search by different keys fast [OK]

Common Mistakes:

• Confusing GSI query with table scan
• Thinking GSI is for deleting or updating keys
• Assuming GSI creates backups

2. Which of the following is the correct way to specify a GSI name in a DynamoDB query using AWS SDK?

easy

A. query({ TableName: 'MyTable', Index: 'MyGSI', KeyConditionExpression: 'GSIKey = :val', ExpressionAttributeValues: { ':val': '123' } })

B. query({ TableName: 'MyTable', KeyName: 'MyGSI', KeyConditionExpression: 'GSIKey = :val', ExpressionAttributeValues: { ':val': '123' } })

C. query({ TableName: 'MyTable', IndexName: 'MyGSI', KeyConditionExpression: 'GSIKey = :val', ExpressionAttributeValues: { ':val': '123' } })

D. query({ TableName: 'MyTable', GSIName: 'MyGSI', KeyConditionExpression: 'GSIKey = :val', ExpressionAttributeValues: { ':val': '123' } })

Solution

1. Step 1: Recall the correct parameter name for GSI in DynamoDB query

   The correct parameter to specify a GSI is 'IndexName'.

2. Step 2: Check the options for correct syntax

   Only query({ TableName: 'MyTable', IndexName: 'MyGSI', KeyConditionExpression: 'GSIKey = :val', ExpressionAttributeValues: { ':val': '123' } }) uses 'IndexName' correctly with other required fields.

3. Final Answer:

   query({ TableName: 'MyTable', IndexName: 'MyGSI', KeyConditionExpression: 'GSIKey = :val', ExpressionAttributeValues: { ':val': '123' } }) -> Option C

4. Quick Check:

   Use IndexName to specify GSI [OK]

Hint: Use 'IndexName' to specify GSI in query params [OK]

Common Mistakes:

• Using 'KeyName' or 'GSIName' instead of 'IndexName'
• Omitting ExpressionAttributeValues
• Wrong key condition syntax

3. Given a DynamoDB table with a GSI named 'StatusIndex' on attribute 'status', what will the following query return?

query({ TableName: 'Orders', IndexName: 'StatusIndex', KeyConditionExpression: 'status = :s', ExpressionAttributeValues: { ':s': 'shipped' } })

medium

A. All orders with status 'shipped' using the GSI

B. All orders regardless of status

C. An error because 'status' is not the primary key

D. No results because GSI queries are not supported

Solution

1. Step 1: Understand the query parameters

   The query uses 'StatusIndex' GSI with a key condition on 'status' = 'shipped'.

2. Step 2: Determine what the query returns

   It returns all items where 'status' attribute equals 'shipped' using the GSI, which is designed for this attribute.

3. Final Answer:

   All orders with status 'shipped' using the GSI -> Option A

4. Quick Check:

   Query GSI by status = shipped returns matching items [OK]

Hint: Query GSI with key condition returns matching items [OK]

Common Mistakes:

• Thinking GSI query returns all items
• Assuming GSI query causes error if attribute not primary key
• Believing GSI queries are unsupported

4. You wrote this DynamoDB query to use a GSI but get an error:

query({ TableName: 'Products', IndexName: 'CategoryIndex', KeyConditionExpression: 'category = :c', ExpressionAttributeValues: { ':c': 'books' } })

What is the most likely cause of the error?

medium

A. You cannot query a GSI in DynamoDB

B. The attribute 'category' is not the partition key of the GSI 'CategoryIndex'

C. ExpressionAttributeValues is missing a value

D. The table name 'Products' is incorrect

Solution

1. Step 1: Understand GSI key requirements

   When querying a GSI, the KeyConditionExpression must use the GSI's partition key attribute.

2. Step 2: Identify the error cause

   If 'category' is not the partition key of 'CategoryIndex', the query will fail with an error.

3. Final Answer:

   The attribute 'category' is not the partition key of the GSI 'CategoryIndex' -> Option B

4. Quick Check:

   GSI query key must be GSI partition key [OK]

Hint: GSI query key must match GSI partition key [OK]

Common Mistakes:

• Using a non-key attribute in KeyConditionExpression
• Assuming table name typo causes this error
• Believing GSI queries are unsupported

5. You have a DynamoDB table 'Employees' with a GSI named 'DeptIndex' where the partition key is 'department' and sort key is 'hireDate'. You want to find all employees in 'Sales' hired after '2023-01-01'. Which query is correct?

hard

A. query({ TableName: 'Employees', IndexName: 'DeptIndex', KeyConditionExpression: 'department = :d AND hireDate > :h', ExpressionAttributeValues: { ':d': 'Sales', ':h': '2023-01-01' } })

B. query({ TableName: 'Employees', IndexName: 'DeptIndex', KeyConditionExpression: 'department = :d OR hireDate > :h', ExpressionAttributeValues: { ':d': 'Sales', ':h': '2023-01-01' } })

C. query({ TableName: 'Employees', KeyConditionExpression: 'department = :d AND hireDate > :h', ExpressionAttributeValues: { ':d': 'Sales', ':h': '2023-01-01' } })

D. query({ TableName: 'Employees', IndexName: 'DeptIndex', FilterExpression: 'department = :d AND hireDate > :h', ExpressionAttributeValues: { ':d': 'Sales', ':h': '2023-01-01' } })

Solution

1. Step 1: Use the GSI name and correct key condition

   Query must specify 'IndexName' as 'DeptIndex' and use KeyConditionExpression with partition key and sort key conditions.

2. Step 2: Check logical operators and expressions

   KeyConditionExpression supports 'AND' for partition key and sort key conditions; 'OR' is invalid here.

3. Step 3: Verify FilterExpression usage

   FilterExpression filters after query; it cannot replace KeyConditionExpression for keys.

4. Final Answer:

   query({ TableName: 'Employees', IndexName: 'DeptIndex', KeyConditionExpression: 'department = :d AND hireDate > :h', ExpressionAttributeValues: { ':d': 'Sales', ':h': '2023-01-01' } }) -> Option A

5. Quick Check:

   Use KeyConditionExpression with AND on GSI keys [OK]

Hint: Use AND in KeyConditionExpression for partition and sort keys [OK]

Common Mistakes:

• Using OR instead of AND in KeyConditionExpression
• Omitting IndexName for GSI query
• Using FilterExpression instead of KeyConditionExpression for keys