6 Semistructured Data

In previous lectures, we’ve worked with relational data, where the schema is well-defined, and each row has a fixed set of attributes or columns. Each attribute’s type is pretty simple (e.g., int), and there is no nesting. But not all data is so well-formed. In recent years, unstructured data has been stored in larger quantities than structured data, due to storage costs decreasing and the ease of logging unstructured data. How do we work with this kind of data?

• Data 101 Course Notes
• 1 SQL
  • 1.1 SQL Review
  • 1.2 String Manipulation
  • 1.3 Truncating Relations
  • 1.4 Creating Tables and Views from Data
  • 1.5 Common Table Expressions (CTE)
  • 1.6 Subqueries
  • 1.7 Window Functions
  • 1.8 DML—Data Manipulation Language
  • 1.9 DDL–Data Definition Language
  • 1.10 Keys & Constraints
• 2 Relational Algebra
  • 2.1 Relational Algebra Overview
  • 2.2 Extended RA Operators
• 3 Query Performance
  • 3.1 Index Selection
  • 3.2 Query Planning
  • 3.3 Rules of Thumb
  • 3.4 PostgreSQL Schemas
• 4 Data Modeling
  • 4.1 Data Models
  • 4.2 Entity Relationship(ER) Diagrams
• 5 Data Preparation
  • 5.1 Structural
  • 5.2 Numerical, Granularity
  • 5.3 Outliers
  • 5.4 Imputation and Entity Resolution
• 6 Semistructured Data
  • 6.1 Semistructured Data
  • 6.2 MongoDB
• 7 Sampling
  • Database Sampling
• 8 Transactions
  • 8.1 Transactions
• 9 Parallel/Distributed Computing
  • Parallel and Distributed Computing
• 10 Data Operations and Pipelines
  • Data Operations and Pipelines
• 11 Business Intelligence (BI)
  • Business Intelligence (BI)
• 12 Graph Databases
  • Property Graphs
  • Triplestores/RDFs
  • Ontologies
• Spreadsheets
  • The Spreadsheet Data Model
• Appendix
  • 12.1 SQL Style Guide
  • 12.2 MongoDB Style Guide