NOSQL DATABASES
Evann De Bailliencourt, Stéphane Hamaili, Amaia Peñagaricano, Aiym Raikhanova
SUMMARY
● INTRODUCTION
○ Facts
○ Timeline
○ Benefits
○ Features
● COMPARISON WITH MONGODB
● HOW TO USE IT
● DEMO
2
FACTS
● RavenDB is ranked among the top 10 Document Databases Worldwide (Source DB-Engines).
● Over 1,000 organizations use RavenDB for their data needs.
● RavenDB was the first NoSQL Database to become Fully Transactional.
● RavenDB has 1.5 million instances of RavenDB running throughout its 37,000 locations.
● RavenDB is Open Source.
3
TIMELINE
● MAY 2010 : RavenDB 1.0 becomes the pioneer Document Database to offer fully transactional.
● JANUARY 2013: RavenDB 2.0 is released.
● JULY 2019: RavenDB Cloud is released.
● DECEMBER 2019: RavenDB Cloud is launched on Google Cloud Platform. Latest stable 4.2.6
● SPRING 2020: RavenDB 5.0 scheduled to include time-series data for IoT applications
4
BENEFITS
● Easy to use
○ Either by developers and non-technical people
● Easy to install
○ Downloading the executable, extracting and running it
● Raven studio
○ Front-end to interact with RavenDB. It’s included with any license, including free community version.
5
FEATURES
● Multi-platform
○ Runs on Windows, Linux, macOS, Docker, Raspberry Pi and others.
● Multiple Clients
○ Can be accessed using the major programming languages in the market, including C#, Java, Node, Python, Ruby and
Go!.
6
FEATURES
● Transactional
○ Is the first non-relational database to achieve ACID across the entire database. Maintain the best of SQL while
boosting your capacity to the next level.
● Management options
○ You can set up a distributed data cluster in minutes. Replicate your database in real time so you can be everywhere
at once, and always available to your users.
● Index Support
○ Indexes in RavenDB are one of the strongest points. Being a NoSQL database, it’s and intelligent solution to
avoiding multiple requests to database by merging multiple tables.
7
FEATURES
● Simple CRUD
○ Especially important to developers. It means easily Creating/Updating/Deleting records, quicker to test and to
release, no migration scripts, simple and powerful api.
● Simple NoSQL Querying
○ RavenDB has powerful ways to do query including geospatial, faceted, full-text and map-reduce operations. But
since most of the time, what you want is very easily available to you via the api and indexes.
8
FEATURES
● In-Memory Database
○ You can use to persist data from your application or better yet, use in your tests so you have real db operations
(in-memory) without hassle of creating mocks and simulating your tests.
● Extensions / Bundles
○ You can extend the database with bundles and extensions. Some come already built for us, others can be created
through what they call Bundles. It’s more of a technical solution but very interesting feature to have available.
9
COMPARISON WITH MONGODB
(Source: https://ravendb.net/articles/ravendb-vs-mongodb-performance-cost-and-complexity) 10
Primary DB Model Document Store Document Store
Rank (Documents Stores) #1 #14
Replication Methods Master-slave replication Multi-master replication
MapReduce yes yes
HOW TO USE IT
11
CREATING
12
CREATING
13
THE DOCUMENT STORE
The Document Store is the main Client API object that establishes the communication between your
client application and the RavenDB cluster.
14
THE SESSION
The session, which is derived from the Document Store, is the primary way your client code interacts
with your RavenDB databases.
15
Querying
Raven Query Language(RQL) allows you to execute all available types of queries and is a part of our JavaScript patching API.
</> RQL
from Orders // select
where Lines.Count > 4 // filter
select Lines[].ProductName as ProductNames, // project
OrderedAt, ShipTo.City
16
Querying. Basics
● Indexes are used by RavenDB to satisfy queries. Each query in RavenDB must be expressed by RQL, our query language. Each query must match an index
in order to return the results. The full query flow is as follows:
1. from index | collection
○ First step. When a query is issued, it locates the appropriate index. If our query specifies that index, the task is simple - use this index. Otherwise, a query
analysis takes place and an auto-index is created.
2. where
○ When we have our index, we scan it for records that match the query predicate.
3. load
○ If a query contains a projection that requires any document loads to be processed, they are done just before projection is executed.
4. select
○ From each record, the server extracts the appropriate fields. It always extracts the id() field (stored by default).
○ If a query is not a projection query, then we load a document from storage. Otherwise, if we stored all requested fields in the index, we use them and
continue. If not, the document is loaded from storage and the missing fields are fetched from it.
○ If a query indicates that projection should be used, then all results that were not filtered out are processed by that projection. Fields defined in the
projection are extracted from the index (if stored).
5. include
○ If any includes are defined, then the results are being traversed to extract the IDs of potential documents to include with the results.
6. Return results.
17
Querying. Filtering
Filtering out data and return records that match a given
condition.
WHERE
from index 'Employees/ByFirstAndLastName'
where FirstName = 'Robert' and LastName =
'King'
WHERE-NESTED/NUMERIC PROPERTY
from Orders where ShipTo.City = 'Albuquerque'
WHERE+ANY
from index 'Order/ByOrderLinesCount'
where Lines_ProductName = 'Teatime Chocolate
Biscuits'
18
WHERE+IN
from Products
where endsWith(Name, 'ra')
WHERE+ContainsAny
from index 'BlogPosts/ByTags' where Tags IN
('Development', 'Research')
Querying. Paging
Paging, or pagination, is the process of
splitting a dataset into pages, reading one
page at a time.
19
Querying. Searching
Use the Search() extension method to perform a full-text search on a particular field. Search() accepts a string containing the
desired search terms separated by spaces. These search terms are matched with the terms in the index being queried.
An index's terms are derived from the values of the documents' textual fields. These values were converted into one or more terms
depending on which Lucene analyzer the index used.
Search
from Users where search(Name, 'John Steve')
Multiple Field
from Users where search(Name, 'Steve') or search(Hobbies, 'sport')
20
Indexes. Types of Indexes
Map indexes contain one (or more) mapping functions that indicate which fields from documents should be indexed. They indicate
which documents can be searched by which fields.
Map-Reduce indexes allow complex aggregations to be performed in a two-step process. First by selecting appropriate records
(using the Map function), then by applying a specified reduce function to these records to produce a smaller set of results.
21
Indexes
Create an index that will map documents from the Employees collection and enable querying by FirstName, LastName, or both.
public class Employees_ByFirstAndLastName : AbstractIndexCreationTask<Employee>
{
public Employees_ByFirstAndLastName()
{
Map = employees => from employee in employees
select new
{
FirstName = employee.FirstName,
LastName = employee.LastName
};
}
}
22
DEMO
23
Thank you for your time.
24