Semi-Structured Data Overview

• The provided text covers the topic of semi-structured data
• It is divided into multiple sections that cover the introduction to generalities and semi-structured data
• The text is intended to provide a comprehensive overview of the subject

Computer Data

• Computer data can be interpreted and used by computers
• Data includes such facts as numbers, words, measurements, observations, or simple descriptions,
• It can be in the form of numbers, texts, images, audio, or videos
• Once collected and organized, data becomes the basis of the computer system

Data Classification by Nature

• There are two types of Data classification based on nature: textual and multimedia
• Text data is written and stored in text format, including text and numbers
• Multimedia data groups different formats, containing text, image, audio, and video

Data Classification by Structure

• There are three types of data structures: structured, unstructured, and semi-structured
• Structured data is predefined and formatted with a precise structure before being placed on physical media
• Unstructured data has no defined structure or schema; is used for reports, text files, comments, opinions on social networks, emails, etc
• Semi-structured data has a format/structure; but is not fixed/rigid

Hypertext

• Hypertext is a document that allows transition from one piece of information to another, using hyperlinks
• Example: a webpage
• A hypertext system contains documents linked together by hyperlinks
• These hyperlinks automatically transition the user to another related document
• Hypertext navigation is a non-linear consultation mode
• Example: a web browser
• Hypertext is not necessarily internet-based
• Examples include local navigation for a web browser, or a PDF reader

Hypermedia

• Hypermedia document allows users to navigate between different sections or to other documents
• Example: YouTube video
• Hypermedia systems are designed to contain and present hypermedia documents to users
• The reader can directly access other related information
• Examples include YouTube and Google Earth

Structured Data

• Structured data uses a predefined and expected format
• It may come from different sources, but common factor is that the fields are fixed

Structured Data Characteristics

• Data is in table form with rows and columns clearly defining data attributes
• A rigid structure is defined before data is populated
• Data of the same attribute (column) are of the same type
• It is easy to search, process, and analyze

Database Reminder

• Review of the basis of databases, including definitions of base de données (BDD(R)), système de gestion de base de données (SGBD (R))
• Covers data modeling of an information system (SI) using E/A or UML, relational models, schemas, primary and foreign keys, normalization
• Physical level includes tables, indexes, and SQL

Semi-Structured Data

• Semi-structured data isn't captured or formatted in a conventional way; that is, data isn't following a tabular model or a relational database because the dataset doesn't have a fixed schema
• The same entry can have multiple structures from the same data source

Semi-Structured Data Representation

• Semi-structured data is represented with the hierarchical model
• Leaves of the model represent the data
• Nodes and links represent the data structure.

Structural Comparisons of Data

• Relational models are used data structures, using rigid schemas defined before data loading
• Updates do not affect the structure of the data, is easy to find
• Data is represented as a flat table, making it difficult to manage missing, multi-valued, or multi-order attributes
• Scaling up a structured BDD schema is very difficult
• Hierarchical models are used in semi-structured data, using flexible and extensible schemas defined implicitly in data
• Updates cause a change in the data structure, with the updates being complicated
• Data is easy to represent complicated, and easy to manage the various attributed values
• Scaling is easy compared to structured data.

Semi-Structured Data Problems

• The web is an important source of semi-structured data
• Its importance and volume have grown with internet development
• One needs to store and manipulate new client data, such as navigation history, cookies, and sessions
• There is a need for adequate tools that can manipulate semi-structured data
• Using SGBD poses problems - as data is completely different from structured data
• Tools and method are needed that will efficiently manage this type of data

Solutions for Semi-Structured Data

• To solve the problem, new technologies include XML (eXtensible Markup Language), JSON (JavaScript Object Notation), CSV (Comma-separated Values), TSV (Tab-Separated Values), the Parquet format, and YAML (Yet Another Markup Language)

Examples of XML and JSON

• Text shows sample representation of the XML and JSON example dataset.

Course Objectives

• The objective of the course is to familiarize students with semi-structured data
• It teaches them to manipulate such data using XML technology
• The module will be organized into four chapters, covering generalities, the XML core, XML galaxies, and XML with BDD