9. Document Oriented Databases

Document-Oriented
Databases in Depth
Dr. Fabio Fumarola

Outline
• Introduction
• What is a Document
• DocumentDBs
• MongoDB
– Data Model
– Indexes
– CRUD
– Scaling
– Pros and Cons
2

Document DB introduction
• Documents are the main concept.
• A Document-oriented database stores and retrieves
documents (XML, JSON, BSON and so on).
• Documents are:
– Self-describing
– Hierarchical tree data structures (maps, collection and
scalar values)
3

What is a Document DB?
• Document databases store documents in the value
part of the key-value store where:
– Documents are indexed using a BTree
– and queried using a JavaScript query engine
4

What is a Document DB?
• Documents have differences in their attributes
• But belongs to the same collection
• This is different from relational databases where columns:
– Stores the same type of values
– Or null
5
{
"name": "Phil",
"age": 26,
"status": "A"
}
{
"name": "Phil",
"age": 26,
"status": "A",
"citiesVisited" : ["Chicago", "LA", "San
Francisco"]
}

RDBMS vs Document DB:
Terminology
6

Document DBs
• MongoDB
• CouchDB
• RethinkDB
• RavenDB
7

Documents: Data Model
• Data has a flexible schema
• This helps in matching document to objects
– Each document can match the fields of a document also if
with different structure
• Data is represented as a map
• Relations can be represented as: references and
embedded documents
10

Documents: Structure
References
11

Documents: Structure Embedded
12

Documents: Write Operations
• Writes are atomic at the document level
– A Denormalized data model facilitates atomic write
operations.
– Normalizing the data over multiple collection would
require multiple write operation that are not atomic.
13

Documents: Growth
• Each time a document is updated the modification
are done changing affected attributes
• Each document has a maximum size of 16MB
• If the document size exceeds MongoDB relocates the
document on disk.
• In MongoDB 3.0 this problem is minimized using the
Power of 2 Sized Allocation
14

Documents: ObjectId
• ObjectId is a 12-byte BSON type, constructed using:
– a 4-byte value representing the seconds since the Unix
epoch,
– a 3-byte machine identifier,
– a 2-byte process id, and
– a 3-byte counter, starting with a random value.
• It is an interesting approach to generate keys
considering that documents are retrieved using
document queries
15

Documents: Indexing
• Indexes allows efficient queries on MongoDB.
• They are used to limit the number of documents to
inspect
• Otherwise, it has to scan every document in a
collection.
• By default MongoDB create indexes only on the _id
field
17

Documents: Indexing
• Indexes are created using B-tree and stores data of
fields ordered by values.
• In addition MongoDB returns sorted results by using
the index.
18

Documents: Index Types
• Single Field
20

• Compound Index: indexed by attributes (left to right)
21

Multikey Index:
•to index content in arrays
22

• Geospatial Index: 2d and 2sphere indexes
• Text Indexes: performs tokenization, stopwords
removal and stemming.
• Hashed Indexes: used to provide an hash based
sharding
23

Queries
• Queries specify criteria, or condition that identify
documents
• A query may include projections to specify the fields
to return.
• It is possible to impose limits, skips and sort orders.
26

Query Interface
27
The same query in SQL

Query Behavior
• All queries in MongoDB address a single collection.
• We impose limits, skips, and sort orders.
• The order of documents returned by a query is not defined
unless you specify a sort().
• Operations that modify existing documents (i.e. updates) use
the same query syntax as queries to select documents to
update.
• In aggregation pipeline, the $match pipeline stage provides
access to MongoDB queries.
28

Cursors
• Each Query like db.collection.find() returns a cursor
• To access the documents, you need to iterate the
cursor.
• By default MongDB closes a cursor
– after 10 minutes of inactivity
– or if the client has exhausted the cursor
31
var myCursor = db.inventory.find().addOption(DBQuery.Option.noTimeout);

Data Modification
• Data modification refers to operations that create,
update, or delete data. In MongoDB.
• These operations modify the data of a single
collection.
• For the update and delete operations, it is possible
to specify the criteria to select the documents to
update or remove.
32

Insert
33
Which corresponds to the SQL query

Update Behavior
• By default, the db.collection.update() method updates a
single document.
• However, with the multi option, update() can update all
documents in a collection.
• If the update() method includes upsert: true and no
documents match the query portion of the update operation,
then the update operation creates a new document.
36

Remove Behavior
• By default, db.collection.remove() method removes
all documents that match its query.
• However, the method can accept a flag to limit the
delete operation to a single document.
38

Write Concern
• There are different levels of guarantee for writes
• When inserts, updates and deletes have a weak write
concern, write operations return quickly.
• In some failure cases, write operations issued with weak write
concerns may not persist.
• With stronger write concerns, clients wait after sending a
write operation for MongoDB to confirm the write
operations.
39

Unacknowledged: Default Write
40

Acknowledged
• This is defined in the drivers
41

Journaled
• Operation are acknowledges only after operations are saved
to the journal.
42

Scaling
• The idea of scaling is to add more node to the cluster
of nodes.
• There are two different context to consider:
– Heavy reads
– Heavy writes
45

Scaling: Heavy Reads
• Scaling here can be achieved by adding more read slaves
• All the reads can be directed to the slaves.
• When a node is added it will sync with the other nodes.
• The advantage of this setting is that we do not need to stop
the cluster.
46
rs.add(“mongo_address:27017”)

Scaling: Heavy Writes
• We can start using the Sharding feature.
• Sharding, or horizontal scaling divides the data set
and distributes the data over multiple servers.
• Each shard is an independent database, and
collectively, the shards make up a single logical
database.
49

Sharding in MongoDB
• Shard: store the data
• Query Routers: interface to
client and direct queries
• Config Server: store
cluster’s metadata.
51

Range Based Sharding
• MongoDB divides the data set into ranges determined by the
shard key values to provide range based partitioning.
52

Hash Based Sharding
• MongoDB computes a hash of a field’s value, and then uses
these hashes to create chunks
53

Performance Comparison
• Range based partitioning supports more efficient range
queries.
• However, range based partitioning can result in an uneven
distribution of data.
• Hash based partitioning, by contrast, ensures an even
distribution of data at the expense of efficient range queries.
54

CouchDB
• Written in Erlang.
• Documents are stored using JSON.
• The query language is in Javascript and supports MapReduce
integration.
• One of its distinguishing features is multi-master replication.
• ACID: It implements a form of Multi-Version Concurrency
Control (MVCC) in order to avoid the need to lock the
database file during writes.
• CouchDB does not guarantees (eventual) consistency to be
able to provide both availability and partition tolerance.
56

CouchDB: Features
• Master-Master Replication - Because of the append-
only style of commits.
• Reliability of the actual data store backing the DB
(Log Files)
• Mobile platform support. CouchDB actually has
installs for iOS and Android.
• HTTP REST JSON interaction only. No binary protocol
57

RethinkDB
• The RethinkDB server is written in C++ and runs on
32-bit and 64-bit Linux systems.
• It is a JSON based database.
• It is characterized by push update queries.
• It supports a MapReduce style API, and geospatial
queries.
58
r.table('tv_shows').insert([{ name: 'Star Trek TNG', episodes: 178 },
{ name: 'Battlestar Galactica', episodes: 75 }])

RavenDB
• It is a DocumentDB written in C#
59

Document Store: Advantages
• Documents are independent units
• Application logic is easier to write. (JSON).
• Schema Free:
– Unstructured data can be stored easily, since a document
contains whatever keys and values the application logic
requires.
– In addition, costly migrations are avoided since the
database does not need to know its information schema in
advance.
60

Suitable Use Cases
• Event Logging: where we need to store different types of
event (order_processed, customer_logged).
• Content Management System: because the schema-free
approach is well suited
• Web analytics or Real-Time Analytics: useful to update
counters, page views and metrics in general.
61

When Not to Use
• Complex Transactions: when you need atomic cross-
document operations, but we can use RavenDB or RethinkDB
• Queries against Varying Aggregate Structure: that is when
the structure of your aggregates vary because of data
continuous data evolutions
62

9. Document Oriented Databases

More Related Content

What's hot

Viewers also liked

Similar to 9. Document Oriented Databases

More from Fabio Fumarola

Recently uploaded

9. Document Oriented Databases

Editor's Notes