This assignment helps you understand what table columns deserve an index through a series of exercises on a database schema. You do not type SQL in this assignment, but rather identify which columns deserve indexes, the type of index it deserves, and why. Correct placement of indexes is absolutely critical to database performance. All but the smallest of databases would have terrible performance if no indexes were assigned to any of its table columns.

An index is a physical construct that serves two purposes – speeding up data retrieval and enforcing uniqueness. Indexes are not modeled in logical entity-relationship diagrams, because indexes do not operate at the logical level of abstraction, as do tables and table columns. While the primary purpose of creating an index is to speed up data access, modern relational DBMS also support a type of index that additionally enforces a uniqueness constraint. A uniqueness constraint can be assigned to a table column directly, or is indirectly assigned to a table column when it is assigned a primary key constraint. While a uniqueness constraint is a logical construct, many modern relational DBMS physically implement this constraint through an index.

You might reasonably ask the question, “Why not simply add indexes to every column in the schema?” After all, then we would not need to concern ourselves with index placement. The primary reason is that while indexes speed up reading from the database, indexes slow down writing to the database. Indexes associated with a table slow down writes to that table, because every time data is added to, modified, or deleted from the table, the indexes referencing the data must be modified. A secondary reason is that indexes increase the size of our database, and that not only affects storage requirements, but also affects database performance since the buffer cache will need to handle