Topic group

Because indexes affect the order in which rows appear, they let users find and update information quickly. To make data entry more efficient, consider these questions:

What is the order in which users expect to see the data? For example, they might expect to see a list of companies in alphabetical order, a list of purchase orders by purchase order number, or a list of invoices in chronological order. Indexes should reflect the expected order of information in a table. If users expect the same information in different sequences, you can create multiple indexes—one for each sequence. For example, in the Orders table, you might want separate indexes for the order number, order date, and customer number.

To find rows in a table, what kind of information might users know already? For example, to locate an invoice, users might already have the invoice number, approximate date of the invoice, or the company that submitted the invoice. To speed up the search process, you might want to create indexes for the most common ways a user looks for information.

What kinds of calculations are users going to perform on data in the table? For example, users might want to calculate the average sale per state or the total sales per month. The word "per" is a clue to an index you might want to create—in the first example, indexing the state field and, in the second example, indexing the sales date field. An index can put similar rows in consecutive order so that users can quickly search for the first row in the series and stop processing after the last row in the series. For example, if users want to calculate the total payments to a vendor, consider creating an index for the vendor number or name.