Version 0.1¹

XML and Databases

The chapter suffers from “XML fever”: the urge to do all things in XML. Unfortunately, XML and relational databases are very different: one is based on tree-structured data, the other on table-structured data; and it’s just confusing to cover both in the same chapter, even more so to pretend that XML files are databases.

The Database Advantage Is?

Everybody knows, I hope, that if you buy a book on Amazon.com, or search on google.com, you’re interacting with a database, not a spreadsheet, on the other end of the connection. Why?

Snyder claims the database advantage over spreadsheets is “metadata.” But what is metadata? Trying to explain it using XML tags just confuses the issue. On p. 472 it seems that metadata is the table schema, which gives the name, data type, and a description of each column in the table. (Did tags do that? No.)

But, is even the table schema what gives the database its advantage? In a way, yes, because having the table strictly defined allows us to apply the relational model (the operations on tables), and yet there’s more to it than that. Here are the significant advantages of an industrial-strength database management system (DBMS) (see Silberschatz, Korth, and Sudarshan, Database System Concepts, chapter 1):

1. The DBMS minimizes redundancy.
2. It allows complex data access through a powerful query language (i.e., operations on tables).
3. It enforces data integrity rules, so we have valid data.
4. It provides transactions.
   • A transaction is a sequence of related changes, which is atomic in that it must either all succeed, or all fail.
5. It enables concurrent access by multiple users.
6. It provides fine-grained security.

Does a desk-top database system, like Microsoft Office Access or OpenOffice/LibreOffice Calc, provide all of these features? Of course not. But it gives us 1–3, which lets us make a start on understanding database concepts. It also provides some user-friendly features which may not be present for multi-user database systems: graphical tools for creating tables, designing queries, and reports.

Table Operations

Operations on tables are applied to one or more tables and produce a table as the result, just like operations on numbers (+, -, etc.) are applied to one or more numbers and produce a number as the result.

1. Select
   • select condition from table
   • Forms a new table from the rows for which condition is true
   • Symbol: σ
2. Project
   • project col, … from table
   • Forms a new table from the specified columns col, …
   • Symbol: π
3. Union
   • table1  ∪  table2
   • Forms a new table with all of the rows of table1 and all of the rows of table2 (but without duplicates)
   • Symbol:  ∪  (Snyder:  + )
4. Difference
   • table1  −  table2
   • Forms a new table from the rows of table1 that are not in table2
   • Symbol:  − 
5. Product
   • table1  ×  table2
   • Forms a new table in which each row of table1 is combined with each row of table2 (and we have all of the columns from both tables)
   • Symbol:  × 
   • Also called “Cartesian product”
6. Join
   • table1  ⋈  table2 on condition
   • Forms a new table in which rows from table1 and table2 are combined when they satisfy the condition
   • Used to link tables together in a relevant way
   • Natural join: the condition specifies that all columns with the same name have equal values in the two tables.
   • There are other variations of this
   • Symbol:  ⋈ 

There are other operations, but these are the principal ones.

These are implemented in different computer languages for querying databases, including SQL and QBE.