NoSQL

When you use RDMS to store(insert) data so all records will be stored into ONE table => ONE file in that server 

When you use Non-RDMS to store(insert) data so each record will be stored into EACH Shard => EACH file in EACH server 

=> Which Shard/File/Server? that is for the router to decide.

1. Routers(Mongos) will sends the query request to all shards to check which shard has the data. Once it receives the response from all shards it sends the result to the client application

2. The methods such as updateMany() and deleteMany() triggers the Broadcast Operations

1. Routers(Mongos) instances use the shard key value to determinate the chunk of data. As above id=1.000.010 in Shard2, and router will point cursor to shard 2 to query data

2. The oeration such as insertOne(), updateOne(), deleteOne()

• With above image you will have three tables: user, order && product and need to know: "How many orders does the user order?

• Query: 

SELECT *

FROM user AS u

JOIN order AS o

ON u.id = o.user_id

JOIN product AS p

ON o.product_id = p.id

WHERE u.name like '%dang%';

• Analyze the query above

  • user table: 2 rows

  • order table:  2 rows

  • product table: 3 rows

=> Total rows need to query: 2 x 2 x 3 = 12 rows

• With above image you will have three tables: user, order && product and need to know: "How many orders does the user order?

• Query: 

db.order.find({"user.name": /dang/});

• Analyze the query above

  • order table: 1 row

=> Total row(s) need to query: 1