S3 Tables

Overview

CloudMock emulates AWS S3 Tables, supporting the full TableBucket → Namespace → Table hierarchy with table policies and metadata location updates. S3 Tables is designed for Apache Iceberg table management.

Supported Operations

Operation	Status	Notes
CreateTableBucket	Supported	Validates bucket name (lowercase, no underscores)
GetTableBucket	Supported	Returns bucket details
ListTableBuckets	Supported	Lists all table buckets
DeleteTableBucket	Supported	Deletes bucket and its tables
CreateNamespace	Supported	Creates namespace within a bucket
GetNamespace	Supported	Returns namespace details
ListNamespaces	Supported	Lists namespaces for a bucket
DeleteNamespace	Supported	Deletes a namespace
CreateTable	Supported	Creates table in bucket/namespace
GetTable	Supported	Returns table details
ListTables	Supported	Lists tables for a bucket
UpdateTableMetadataLocation	Supported	Updates Iceberg metadata location
DeleteTable	Supported	Deletes a table
GetTablePolicy	Supported	Returns table resource policy
PutTablePolicy	Supported	Sets table policy (validates JSON)
DeleteTablePolicy	Supported	Deletes table policy

Resource Hierarchy

TableBucket (arn:aws:s3tables:{region}:{account}:bucket/{name})
  └── Namespace (analytics, ml, etc.)
       └── Table (orders, customers, etc.)
            └── TablePolicy (IAM resource policy)

Quick Start

import boto3

client = boto3.client('s3tables',
    endpoint_url='http://localhost:4566',
    region_name='us-east-1',
    aws_access_key_id='test',
    aws_secret_access_key='test')

# Create table bucket
bucket = client.create_table_bucket(name='my-lakehouse')
bucket_arn = bucket['arn']

# Create namespace
client.create_namespace(
    tableBucketARN=bucket_arn,
    namespace=['analytics'],
)

# Create Iceberg table
client.create_table(
    tableBucketARN=bucket_arn,
    namespace='analytics',
    name='events',
    format='ICEBERG',
)

# Update metadata location after Iceberg write
client.update_table_metadata_location(
    tableBucketARN=bucket_arn,
    namespace='analytics',
    name='events',
    metadataLocation='s3://bucket/warehouse/analytics/events/metadata/v1.metadata.json',
    versionToken='old-token',
)

Configuration

services:
  s3tables:
    enabled: true

Known Differences from AWS

• No actual Iceberg table files are created or managed
• MetadataLocation updates store the value but do not validate S3 paths
• Namespace and table ARNs are derived from bucket ARN