The Index Density Report displays information about non-unique indexes. Indexes should only be created on fields where the corresponding data exhibits a high degree of uniqueness. As a rule of thumb, 90 to 95 % of the rows should contain unique values. A possible exception to this rule are indexes on foreign keys when the volume of data is high.
Non-essential indexes increase insert and delete overhead and may confuse the query optimizer resulting in less than optimal performance.
This report lists the non-unique indexes for each table in the current MS-Access database. It shows the number of unique (or distinct) rows, the total number of rows in the table, and percentage of unique rows for the index in the table. It also shows which indexes are based on foreign keys.
In this example, the CompanyName and PostalCode are good candidates for indexes, however the Region and City indexes are probably poor choices.
In the Order Details table, the OrderID is probably not a good index candidate even though it is based on a foreign key since the volume of data is relatively small.