Skip to content

Latest commit

 

History

History
164 lines (105 loc) · 5.17 KB

mongodb.md

File metadata and controls

164 lines (105 loc) · 5.17 KB
Raw

MongoDB

Study notes related to MongoDB storage

Overview

MongoDB is document-oriented NoSQL storage system with flexible quering mechanism. Being that said, it mean that MongoDB do not store data in old fashioned table oriented manner. Instead it store data in Documents with flexible schemas (each document can have uniq set of fields) and groups them in Collections that related to certain database. Mongo shell use JavaScript language for perfoming query operations.

Usage

with Docker

# some other Mongo images can be used
$ docker -it mongo sh

Mongo has cli tool mongo for quering database. Mongo default port 27017

Basic mongo usage

MongoDB interactive shell client.

  • Connect to a database:

    mongo database

  • Connect to a database running on a given host on a given port:

    mongo --host host --port port database

  • Connect to a database with a given username; user will be prompted for password:

    mongo --username username database --password

  • Evaluate a javascript expression on the database:

    mongo --eval 'JSON.stringify(db.foo.findOne())' database

Mongo collection operations

  • Find each document in collection
db.collection.find({})

As an optional second parameter find accept projection. Projection is a way of enabling or disabling certain fields in final data.

So for example, in such case we find all records and disable _id field

db.collection.find({}, {'_id': 0})
  • Insert document into collection

Note! use insert_one or insert_many, pure insert is supported but will be deprecated soon

// Insert single
db.collection.insert_one({ 'name': 'Oleh' })

//Insert many 
// Everything is ok due to flexible schema
db.collection.insert_many({'name': 'test'}, {'surname': 'test'})
  • Update document
db.collection.update(
  { '_id': ObjectId('some object id') }, 
  { '$set': {'data': 'updated_data'} }
)
  • Using operators with query
db.collection.update(
  { 'age': { '$gt': 21 } }, 
  { '$set': { 'data': 'some test data' } }
)
  • Updating multiple documents
db.collection.update(
  { 'age': { '$gt': 21 } }, 
  { '$set': { 'data': 'some test data' } },
  { 'multi': true }
)
  • Using upserts

Upserts useful for matching document and if it exists updating it or creating new one

db.collection.update(
  { 'name': 'Oleh' }, 
  { '$set': { 'data': 'some test data' } },
  { 'upsert': true }
)
  • Remove document from collection
db.collection.delete_one({ '_id': ObjectId('some_id') })

More operationg can be found in docs

Data Modeling

data modeling in NoSQL storages a bit different from traditional SQl based dbs.

Most common pitfall that Mongo don't have join mechanism so instead of performing joins of collections you just create two separete queries for each piece of data.

Mongo allows storing of nested structures, one document's size is limited to 16mb.

When to use Mongo

  • Flexible schema
  • Logs
  • Storing nested structures ( eg. Three.js json models )
  • Geospital data. Mongo support geoJSON and have powerful operators for it.
  • more ?

Data Aggregation

Mongo has two powerful mechanism for data aggregation - aggregate function - MapReduce functions

Replication

TBD

Sharding

TDB

Mongo clients

  • motor async Python3 client for asyncio / Tornado

Resources

List of useful study resources