In this chapter, you learned the following:
Most large applications store information in relational databases. A database is made up of tables, each of which stores logically related information. A table has one or more columns—each of which has a name and a type—and zero or more rows, or records. In most tables, each row can be identified by a unique key, which consists of one or more of the values in the row.
Commands to put data into databases, or to get data out, can be written in a specialized language called SQL.
SQL commands can be sent to databases interactively from GUIs or command-line tools—but for larger jobs, it is more common to write programs that create SQL and process the results.
Changes made to a database don’t actually take effect until they are committed. This ensures that if two or more programs are working with a database at the same time, it will always be in a consistent state. However, it also means that operations in one program can fail because of something that another program is doing.
SQL queries must specify the table(s) and column(s) that values are to be taken from. They may also specify Boolean conditions those values must satisfy and the ordering of results.
Simple queries work on one row at a time, but programs can join tables to combine values from different rows. Queries can also group and aggregate rows to calculate sums, averages, and other values.
Databases can use the special value NULL to represent missing information. However, it must be used with caution, since operations on NULL values don’t behave in the same way that operations on “real” values do.