Scribd
Upload
94 views3 pages

Dynamo DB Cheat Sheet: Partitions - 10% Rule

The document provides tips for structuring data in DynamoDB. It recommends: 1) Using partition keys to distribute data across partitions, keeping each partition below 10% of the total data to avoid throttling. 2) Employing range keys to enable querying on ranges of values rather than exact matches only. 3) Creating compound sort keys that concatenate multiple attributes to enable hierarchical grouping and querying of related data. 4) Storing multiple data types for a single entity in one table using composite keys that identify the data type and item.

Uploaded by

Miguel Lorenzo Singian
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
94 views3 pages

Dynamo DB Cheat Sheet: Partitions - 10% Rule

The document provides tips for structuring data in DynamoDB. It recommends: 1) Using partition keys to distribute data across partitions, keeping each partition below 10% of the total data to avoid throttling. 2) Employing range keys to enable querying on ranges of values rather than exact matches only. 3) Creating compound sort keys that concatenate multiple attributes to enable hierarchical grouping and querying of related data. 4) Storing multiple data types for a single entity in one table using composite keys that identify the data type and item.

Uploaded by

Miguel Lorenzo Singian
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Dynamo DB Cheat Sheet

Partitions - 10% Rule

Use Range Keys to make querying data easier

Using Compound Sort Keys for grouping data

One Table - Multiple stored datatypes

Partitions - 10% Rule


Never have a partition that contains more than 10% of the table data.

When Dynamo is storing your data it groups it all by the partition key. Multiple partitions can be
stored on a single SSD, but one partition can never be split across multiple SSDs. This means
that you want to make sure that you always create small partition keys. You don’t want to have
one partition, and therefore one SSD, getting hit with most of the requests and becoming
throttled.

Use Range Keys to make querying data easier


As well as a Partition Key you can have a Range Key. This key is different from a partition key
in that you can query on more than an exact match.
Condition Example

= rangeKey = 1598942275

< rangeKey < 1598942275

<= rangeKey <= 1598942275

> rangeKey > 1598942275

>= rangeKey >= 1598942275

BETWEEN rangeKey BETWEEN 1598242275 AND 1598942275

begins_with() begins_with( rangeKey, ‘15983’ )

contains() contains(rangeKey, ‘82422’)


Using Compound Sort Keys for grouping data
Sort keys can be used for more than querying on a date. It can be used for hierarchical
grouping. This is where you chain increasingly small categories into the sort key. This is easier
explained with examples.

If you have a table containing store information for stores that you sell your products to.
Partition Key Store Brand

Sort Key {country}#{state}#{city}#{town/area}#{ID}

This way you can query make these queries:


Request Query

All Target in New York PartitionKey = Target AND begins_with(SortKey, USA#NY)

All Wallmart in Seattle PartitionKey = Wallmart AND begins_with(SortKey,


USA#WA#seattle)

All Wallmart in Beacon Hill, Seattle PartitionKey = Wallmart AND begins_with(SortKey,


USA#WA#seattle#beacon_hill)

A specific Target that is in Roxhil, PartitionKey = Wallmart AND SortKey =


Seattle USA#WA#seattle#roxhill#02349283

Using this you are able to get groups of data but still get a specific item of information.
One Table - Multiple stored datatypes
In another way to use compound keys are to use them to store multiple types of data in a single
table.
An example would be if you are storing data about a customer, traditionally you would have a
table for each type of data you are storing. A table for addresses, a table for orders, etc. With
Dynamo, you can have this all in one table if you create your Sort key correctly.

Partition Key Customer ID

Sort Key {DataType}#{date}#{ID}

This would end up with a table like this:


Partition Key Sort Key Data… (more fields)

12345 address#0321 ....

12345 order#1593115224#10352 ....

12345 order#1597115224#10367 ...

12345 order#1599115224#10394 ...

12345 complaint#1597815224#0031 ...

12345 complaint#1599995224#0031 ...

.... ... ...

You could then do queries like this:


Request Query Result

Get all orders for user 12345 PartitionKey = 12345 AND begins_with(SortKey, 3 orders
order)

Get all orders for 12345 in PartitionKey = 12345 AND SortKey BETWEEN 1 order
August 1596236400 AND 1598914800

Get order 10367 PartitionKey = 12345 AND SortKey = 10367 1 order

Get all details for user 12345 PartitionKey = 12345 1 address, 3


orders, 2
complaints

You might also like