Data choreography in mongo
0 likes347 views
This document provides an overview of data modeling principles for MongoDB, a NoSQL database. It discusses: 1) Schema design principles for MongoDB including supporting ACID at the document level, embedding data vs references, and how schema impacts queries, joins, and sharding. 2) An example data model for a corporate HRMS using embedded documents to represent relationships between entities like employees, projects, roles and addresses. 3) A note that part 2 will provide more examples and techniques for design, migration, performance tuning and map-reduce in MongoDB.
![An example document design for a
corporate HRMS entities
Entities – Employee, Address, Role, Project …..
Employee Master Projects Master Employee {
address: {[
Role Master …….
Emp-Projects ]},
role: {
Emp-Role
r_id: xxxx,
…….
Addresses Master },
project: { [ prj-id:yyyy,
Employee-Addresses …….
]}
}
……
}
RDBMS World Mongo DB World](https://image.slidesharecdn.com/datachoreographyinmongo-121213003226-phpapp02/85/Data-choreography-in-mongo-5-320.jpg)
Data choreography in mongo
- 1. Basic Data Choreography (Modeling) in MongoDB! Part 1 A first-munch by VulcanMinds
- 2. By the Way…What is a Data Model?
- 3. Data Modeling with NOSQL Relational Data Modeling NoSQL Data Modeling
- 4. Schema Design Principles Mongo supports ACID at a document level only. Schema should be driven by the data entities and usage patterns ( use cases or query patterns) Embed everything Vs No-Embed ( atomic like an RDBMS) is a well thought-out tradeoff! Application-side /runtime joins Vs design-time joins… Mongo is happier with design-time joins. ‘Frequent updates to entities’ is a good symptom to move it to separate independent entities. Design such a way to easily lend itself to map-reduce. BTW will your business process need to map-reduce? Schema design also affects the clustering of MongoDB instances with shard key choices and run-time query and I/O performance metrics ( more about replica sets and
- 5. An example document design for a corporate HRMS entities Entities – Employee, Address, Role, Project ….. Employee Master Projects Master Employee { address: {[ Role Master ……. Emp-Projects ]}, role: { Emp-Role r_id: xxxx, ……. Addresses Master }, project: { [ prj-id:yyyy, Employee-Addresses ……. ]} } …… } RDBMS World Mongo DB World
- 6. Pros and Cons Relational Data Modeling NoSQL Data Modeling
- 7. Coming soon in Part 2! • More complete examples and techniques for – design, migration, performance –tuning and map-reduce! Till then ….. See you!