How to Use Condition Expression in DynamoDB for Safe Writes

In DynamoDB, use a ConditionExpression to specify conditions that must be true for an operation like PutItem or UpdateItem to succeed. This helps prevent overwriting existing data or updating only when certain criteria are met. The condition uses attribute names and values with operators like =, attribute_exists(), or attribute_not_exists().

Syntax

A ConditionExpression is a string that defines a condition for DynamoDB operations. It uses attribute names, placeholders, and functions to check item attributes.

• Attribute names: Refer to item fields, often with placeholders like #name.
• Attribute values: Use placeholders like :val to represent values.
• Functions: Common ones include attribute_exists() and attribute_not_exists().
• Operators: Use =, <, >, AND, OR to build conditions.

Example syntax:

plaintext

ConditionExpression = "attribute_not_exists(#pk) AND #age >= :minAge"
ExpressionAttributeNames = {"#pk": "UserId", "#age": "Age"}
ExpressionAttributeValues = {":minAge": {"N": "18"}}

Example

This example shows how to use a condition expression with PutItem to add a user only if the user ID does not already exist.

python

import boto3

# Create DynamoDB client
client = boto3.client('dynamodb')

# Define item to add
item = {
    'UserId': {'S': 'user123'},
    'Name': {'S': 'Alice'},
    'Age': {'N': '25'}
}

try:
    response = client.put_item(
        TableName='Users',
        Item=item,
        ConditionExpression='attribute_not_exists(UserId)'
    )
    print('Item added successfully')
except client.exceptions.ConditionalCheckFailedException:
    print('Item with this UserId already exists')

Output

Item added successfully

Common Pitfalls

Common mistakes when using condition expressions include:

• Not using ExpressionAttributeNames when attribute names are reserved words or contain special characters.
• Forgetting to handle ConditionalCheckFailedException, which occurs if the condition is false.
• Using incorrect syntax or operators in the condition expression.
• Confusing attribute existence functions like attribute_exists() and attribute_not_exists().

Example of a wrong and right way:

python

# Wrong: Using reserved word 'Name' directly
client.put_item(
    TableName='Users',
    Item=item,
    ConditionExpression='attribute_not_exists(Name)'
)

# Right: Using ExpressionAttributeNames to avoid reserved word conflict
client.put_item(
    TableName='Users',
    Item=item,
    ConditionExpression='attribute_not_exists(#nm)',
    ExpressionAttributeNames={'#nm': 'Name'}
)

Quick Reference

Function/Operator	Description	Example
attribute_exists(path)	True if attribute exists	attribute_exists(UserId)
attribute_not_exists(path)	True if attribute does not exist	attribute_not_exists(UserId)
=	Equals operator	#age = :val
<>	Not equals operator	#status <> :val
AND	Logical AND	attribute_exists(UserId) AND #age > :minAge
OR	Logical OR	attribute_not_exists(UserId) OR #age < :maxAge

Key Takeaways

Use ConditionExpression to control when DynamoDB operations succeed based on item attributes.

Always handle ConditionalCheckFailedException to catch failed conditions.

Use ExpressionAttributeNames to avoid conflicts with reserved words.

Common functions include attribute_exists() and attribute_not_exists() for checking attribute presence.

Combine conditions with AND, OR, and comparison operators for precise control.