nosql-expert

name	nosql-expert
description	Models Cassandra, ScyllaDB, and DynamoDB access patterns as tables: high-cardinality partition keys, clustering or sort keys, single-table adjacency lists, GSIs, and duplicated lookup tables. Trigger on hot partitions, ALLOW FILTERING, or DynamoDB single-table design. Never apply MongoDB document schemas or SQL join thinking to these stores.
version	1.0.1
risk	unknown
source	community
date_added	2026-02-27

NoSQL Expert Patterns (Cassandra, ScyllaDB & DynamoDB)

Overview

This skill provides professional mental models and design patterns for distributed wide-column and key-value stores (Apache Cassandra, ScyllaDB, and Amazon DynamoDB).

Unlike SQL (where you model data entities), or document stores (like MongoDB), these distributed systems require you to model your queries first. You cannot "add a query later" without migration or creating a new table/index.

The Golden Rule: In SQL, you design the data model to answer any query. In NoSQL, you design the data model to answer specific queries efficiently.

The Mental Shift: SQL vs. Distributed NoSQL

Feature	SQL (Relational)	Distributed NoSQL (Cassandra/DynamoDB)
Data modeling	Model Entities + Relationships	Model Queries (Access Patterns)
Joins	CPU-intensive, at read time	Pre-computed (Denormalized) at write time
Storage cost	Expensive (minimize duplication)	Cheap (duplicate data for read speed)
Consistency	ACID (Strong)	BASE (Eventual) / Tunable
Scalability	Vertical (Bigger machine)	Horizontal (More nodes/shards)

When to Use

• Designing for Scale: Moving beyond simple single-node databases to distributed clusters.
• Technology Selection: Evaluating or using Cassandra, ScyllaDB, or DynamoDB.
• Schema Modeling: Designing tables, partition keys, sort keys, or single-table layouts.
• Performance Tuning: Troubleshooting "hot partitions", high latency, or uneven traffic in existing NoSQL systems.
• Microservices: Implementing "database-per-service" patterns where highly optimized reads are required.

Prerequisites

• Target database identified (Cassandra/ScyllaDB or DynamoDB).
• A complete list of required access patterns (queries) before table design.
• Understanding of expected read/write volume and cardinality of candidate partition keys.
• If running locally on Windows (PowerShell), ensure cqlsh or AWS CLI v2 is installed and configured with placeholder credentials (YOUR_KEY).

Procedure

1. Query-First Modeling (Access Patterns)

You typically cannot "add a query later" without migration or creating a new table/index.

1. List all Entities (User, Order, Product).
2. List all Access Patterns ("Get User by Email", "Get Orders by User sorted by Date").
3. Design Table(s) specifically to serve those patterns with a single lookup.
4. Validate that every access pattern maps to exactly one table or index.

2. Choose the Partition Key (PK)

Data is distributed across physical nodes based on the Partition Key (PK).

• Goal: Even distribution of data and traffic.
• Anti-Pattern: Using a low-cardinality PK (e.g., status="active" or gender="m") creates Hot Partitions, limiting throughput to a single node's capacity.
• Best Practice: Use high-cardinality keys (User IDs, Device IDs, Composite Keys).
• Split Partition Risk: For any single partition (e.g., a single user's orders), will it grow indefinitely? If a partition may exceed 10GB, shard it (e.g., USER#123#2024-01).

3. Choose Clustering / Sort Keys

Within a partition, data is sorted on disk by the Clustering Key (Cassandra) or Sort Key (DynamoDB).

• Enables efficient Range Queries (e.g., WHERE user_id=X AND date > Y).
• Pre-sorts data for specific retrieval requirements.

4. Single-Table Design (Adjacency Lists)

Primary use: DynamoDB (but concepts apply elsewhere)

Storing multiple entity types in one table to enable pre-joined reads.

PK (Partition)	SK (Sort)	Data Fields...
USER#123	PROFILE	{ name: "Ian", email: "..." }
USER#123	ORDER#998	{ total: 50.00, status: "shipped" }
USER#123	ORDER#999	{ total: 12.00, status: "pending" }

• Query: PK="USER#123"
• Result: Fetches User Profile AND all Orders in one network request.

5. Denormalization & Duplication

Store the same data in multiple tables to serve different query patterns.

• Table A: users_by_id (PK: uuid)
• Table B: users_by_email (PK: email)

Trade-off: You must manage data consistency across tables (often using eventual consistency or batch writes).

6. Apache Cassandra / ScyllaDB Specifics

• Primary Key Structure: ((Partition Key), Clustering Columns)
• No Joins, No Aggregates: Do not try to JOIN or GROUP BY. Pre-calculate aggregates in a separate counter table.
• Avoid ALLOW FILTERING: If you see this in production, your data model is wrong. It implies a full cluster scan.
• Writes are Cheap: Inserts and Updates are just appends to the LSM tree. Don't worry about write volume as much as read efficiency.
• Tombstones: Deletes are expensive markers. Avoid high-velocity delete patterns (like queues) in standard tables.

7. AWS DynamoDB Specifics

• GSI (Global Secondary Index): Use GSIs to create alternative views of your data (e.g., "Search Orders by Date" instead of by User). GSIs are eventually consistent.
• LSI (Local Secondary Index): Sorts data differently within the same partition. Must be created at table creation time.
• WCU / RCU: Understand capacity modes. Single-table design helps optimize consumed capacity units.
• TTL: Use Time-To-Live attributes to automatically expire old data (free delete) without creating tombstones.

Examples

Cassandra: Users by Email Lookup

CREATE TABLE users_by_email (
  email text,
  user_id uuid,
  name text,
  created_at timestamp,
  PRIMARY KEY (email)
);

Query: SELECT * FROM users_by_email WHERE email = '[e-mail]';

DynamoDB: Single-Table User + Orders

PK: USER#123          SK: PROFILE         -> { name, email }
PK: USER#123          SK: ORDER#2024-001 -> { total, status }
PK: USER#123          SK: ORDER#2024-002 -> { total, status }

Query: Query PK=USER#123 returns the profile and all orders in one request.

Pitfalls

• ❌ Scatter-Gather: Querying all partitions to find one item (Scan).
• ❌ Hot Keys: Putting all "Monday" data into one partition.
• ❌ Relational Modeling: Creating Author and Book tables and trying to join them in code. Instead, embed Book summaries in Author, or duplicate Author info in Books.
• ❌ Low-Cardinality PK: status, gender, or country as a partition key creates hot partitions.
• ❌ Unbounded Partitions: A single user's orders growing forever without sharding.
• ❌ ALLOW FILTERING in Cassandra: Indicates a broken data model requiring a full cluster scan.
• ❌ High-Velocity Deletes: Creates tombstones that degrade read performance in Cassandra.
• ❌ Forgetting GSI Consistency: GSIs are eventually consistent; do not use them for strong-consistency requirements.

Verification

Before finalizing your NoSQL schema, run through this checklist:

• Access Pattern Coverage: Does every query pattern map to a specific table or index?
• Cardinality Check: Does the Partition Key have enough unique values to spread traffic evenly?
• Split Partition Risk: For any single partition (e.g., a single user's orders), will it grow indefinitely? (If > 10GB, shard the partition, e.g., USER#123#2024-01).
• Consistency Requirement: Can the application tolerate eventual consistency for this read pattern?
• No Scans: Are all production queries SELECT/Query by key, not Scan?
• No ALLOW FILTERING: Confirm no Cassandra queries use ALLOW FILTERING.

Checkable Commands

Cassandra/ScyllaDB (Windows PowerShell, cqlsh on PATH):

cqlsh localhost 9042 -e "DESCRIBE TABLE keyspace.users_by_email;"
cqlsh localhost 9042 -e "SELECT COUNT(*) FROM keyspace.users_by_email WHERE email='[e-mail]';"

DynamoDB (AWS CLI v2, placeholder credentials):

aws dynamodb describe-table --table-name UsersOrders --endpoint-url http://localhost:8000
aws dynamodb query --table-name UsersOrders --key-condition-expression "PK = :pk" --expression-attribute-values '{":pk":{"S":"USER#123"}}' --endpoint-url http://localhost:8000

Limitations

• Use this skill only when the task clearly matches the scope described above.
• Do not treat the output as a substitute for environment-specific validation, testing, or expert review.
• Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.