23/10/2022
BUCKET PATTERN
Patterns in NOSQL Data Modeling
1. Attribute Pattern 7. Tree Patterns
2. Extended Reference Pattern 8. Polymorphic Pattern
3. Subset Pattern 9. Other Patterns
4. Computed Pattern
5. Bucket Pattern
6. Schema Versioning Pattern
1
23/10/2022
Bucket Pattern
Too Many or Too Big
˗ One document per piece of
information per device
˗ One document will all the
information for a device
2
23/10/2022
Just the Right Amount
One document per device per day One document per device per hour
{ “device_id”: 00123456, { “device_id”: 00123456,
“type”: “2A” “type”: “2A”
“date”: ISODate(“2018-03-02”) “date”: ISODate(“2018-03-02T13”)
“temp”: [ [20.0, 20.1, 20.2, …], “temp”: { 1: 20.0, 2: 20.1, 3: 20.2, …}
[22.1, 22.1, 22.0, …] }
.... ] }
{ “device_id”: 00123456, { “device_id”: 00123456,
“type”: “2A” “type”: “2A”
“date”: ISODate(“2018-03-03”) “date”: ISODate(“2018-03-02T14”)
“temp”: [ [20.1, 20.2, 20.3, …], “temp”: { 1: 22.1, 2: 22.1, 3: 20.0, …}
[22.4, 22.4, 22.3, …] }
....] }
Collaboration Platform Example
˗ Many channels
{
˗ Design alternatives “channel_id”, “13579246”,
“name”: “mongodb_tech”,
One document per message “date”: ISODatte(“2018-03-02T03:14:16”),
“message”:”<Daniel joined the group>”
}
3
23/10/2022
Collaboration Platform Example
˗ Many channels
˗ Design alternatives
{
One document per message “channel_id”, “13579246”,
One document for all “name”: “mongodb_tech”,
“message”: [
massages in a channel “<Channel created>”,
“< Nathan joined the group >”,
“< Daniel joined the group >”,
“Nathan: welcome Daniel”,
“Daniel: Thanks!”
…
}
Collaboration Platform Example
˗ Many channels
˗ Design alternatives
One document per message
One document for all
massages in a channel
One document per channel,
per day
4
23/10/2022
Row-based vs Column-based approach
Column Oriented Data
5
23/10/2022
Gotchas with Buckets
˗ Random insertions or deletions in buckets
˗ Difficult to sort across buckets
˗ Ad hoc queries may be more complex, again across buckets
˗ Works best when the “complexity” is hidden through the
application code
Bucket Pattern
6
23/10/2022
Summary
˗ Alternative to fully
embedding or fully linking a
1-to-Many relationship
˗ Advanced pattern that
requires a good understand
of the workload
Lab: Apply the Bucket Pattern
7
23/10/2022
Patterns in NOSQL Data Modeling
1. Attribute Pattern 7. Tree Patterns
2. Extended Reference Pattern 8. Polymorphic Pattern
3. Subset Pattern 9. Other Patterns
4. Computed Pattern
5. Bucket Pattern
6. Schema Versioning Pattern
SCHEMA VERSIONING PATTERN
8
23/10/2022
Alter table nightmares - Relational Database
MongoDB Database
9
23/10/2022
Updating a Relational Database Schema
˗ Need time to update the Data
˗ Usually done by stopping the Application
˗ Hard to revert if something goes wrong
NoSQL Database - MongoDB
10
23/10/2022
Schema versioning pattern
Application Lifecycle
11
23/10/2022
Document Lifecycle
Timeline of the Migration
12
23/10/2022
Schema Versioning Pattern
Summary
13
23/10/2022
Lab: Apply the Schema Versioning Pattern
˗ Problem:
Which of the following scenarios are best suited for applying the Schema Versioning Pattern?
˗ Scenario A:
Your team was assigned to upgrade the current schema with additional fields and transforming the
type of different fields without bringing the system down for this upgrade. However, all documents
need to be updated to the new shape quickly.
˗ Scenario B:
The performance of your application became suboptimal over time. Your team has identified that
the most commonly used collection could profit from embedding additional information from other
collections using the Subset and Computed Patterns. All documents in the commonly-used
collection will need to undergo this modification. If possible, you would like the transition to be
done without downtime.
˗ Scenario C:
Your company was bought by its slightly more successful competitor. Thankfully both your and
your new owner's applications are flexible enough to handle both document shapes well. You do
not have to modify the application or your document shape, but due to the merger, you have to
keep documents with different structures in the same collection.
Patterns in NOSQL Data Modeling
1. Attribute Pattern 7. Tree Patterns
2. Extended Reference Pattern 8. Polymorphic Pattern
3. Subset Pattern 9. Other Patterns
4. Computed Pattern
5. Bucket Pattern
6. Schema Versioning Pattern
14
23/10/2022
TREE PATTERNS
Tree Patterns
15
23/10/2022
Tree Patterns
Tree Patterns
16
23/10/2022
Tree Patterns
Tree Patterns
17
23/10/2022
Tree Patterns
˗ Who are the ancestors of node X?
˗ Who report to Y?
˗ Find all nodes that are under Z?
˗ Change all categories under N to under P
Tree Patterns
18
23/10/2022
Model Tree Structures
˗ Parent References
˗ Child References
˗ Array of Ancestors
˗ Materialized Paths
Parent References
{
Name: “Office”,
Parent: “Swag”,
…
}
19
23/10/2022
Parent References
{
Name: “Office”,
Parent: “Swag”,
…
}
Parent References
{
Name: “Office”,
Parent: “Swag”,
…
}
20
23/10/2022
Parent References
˗ Who are the ancestors of node X? (!)
˗ Who report to Y? ()
˗ Find all nodes that are under Z? (!)
˗ Change all categories under N to under P ()
Child References
{
name: “Office”,
Children“:[“Books”,
“Electronics”, “Stickers”]
…
}
21
23/10/2022
Child References
˗ Who are the ancestors of node X? (!)
˗ Who report to Y? (!)
˗ Find all nodes that are under Z? ()
˗ Change all categories under N to under P (!)
Array of Ancestors
{
Name: “Books”,
Parent: [“Swag”, “Office”]
…
}
22
23/10/2022
Array of Ancestors
˗ Who are the ancestors of node X? ()
˗ Who report to Y? ()
˗ Find all nodes that are under Z? ()
˗ Change all categories under N to under P (!)
Array of Ancestors
˗ Who are the ancestors of node X? ()
˗ Who report to Y? ()
˗ Find all nodes that are under Z? (!)
˗ Change all categories under N to under P (!)
23
23/10/2022
Materialized Paths
{
Name: “Books”,
Parent: “.Swag.Office”]
…
}
Materialized Paths
{
Name: “Books”,
Parent: “.Swag.Office”]
…
}
24
23/10/2022
Materialized Paths
˗ Who are the ancestors of node X? ()
˗ Who report to Y? (!)
˗ Find all nodes that are under Z? (!)
˗ Change all categories under N to under P (!)
MongoMark Solution: Ancestor Array +
Parent Reference
{
“_id” : 8,
“name” : “Umbrellas”,
“parent” : [ “Swag”, “Fashion”]
}
25
23/10/2022
MongoMark Solution: Ancestor Array +
Parent Reference
{
“_id” : 8,
“name” : “Umbrellas”,
“parent” : [ “Swag”, “Fashion”]
}
Tree Patterns
26
23/10/2022
Recap
˗ Documents are good data structure to represent hierarchical
data
˗ Several different patterns to represent trees
˗ Focus on the most common queries and operations to select
the most effective tree pattern
Patterns in NOSQL Data Modeling
1. Attribute Pattern 7. Tree Patterns
2. Extended Reference Pattern 8. Polymorphic Pattern
3. Subset Pattern 9. Other Patterns
4. Computed Pattern
5. Bucket Pattern
6. Schema Versioning Pattern
27
23/10/2022
POLYMORPHIC PATTERN
Differences in Objects
28
23/10/2022
More in Common than Differences
Vehicles
29
23/10/2022
Products
Polymorphic Subdocuments
30
23/10/2022
Single View
Single View Example
31
23/10/2022
Polymorphism in the Schema Versioning
Pattern
Polymorphic Patterns
32
23/10/2022
Summary
˗ Basic Pattern
˗ Base of many other Patterns
Patterns in NOSQL Data Modeling
1. Attribute Pattern 7. Tree Patterns
2. Extended Reference Pattern 8. Polymorphic Pattern
3. Subset Pattern 9. Other Patterns
4. Computed Pattern
5. Bucket Pattern
6. Schema Versioning Pattern
33
23/10/2022
Other Patterns
˗ Approximation Pattern
˗ Outlier Pattern
Approximation Pattern
34
23/10/2022
Approximation Pattern
Outlier Pattern
35
23/10/2022
Movies Site Solution
Outlier Pattern
36
23/10/2022
Recap
These are some other notable patterns
˗ Approximation Pattern
Avoiding performing an operation too often
˗ Outlier Pattern
Keeping the focus on the most frequent use cases
Summary of Patterns
37
23/10/2022
Summary of Patterns
˗ Pattern are Powerful Transforms for your Schema
˗ Provide a common Language for your Team
˗ More Predictable Methodology
38
