Person: Dimensions are attributes or characteristics of your data. Specifically, each column in a database table is a dimension in Looker. For example, this is a basic 12-row data set that describes the fruit basket in the image on the left. Each piece of fruit is represented as a row in the table, with each column describing a different attribute or aspect of the fruit. Each fruit has a type and a color, either is or is not round and has a price per pound, weight and unit price. In Looker, all of these columns are dimensions. Using this data set, you can ask a question like, "What types of fruit are in the basket?" What is the dimension you need here? Fruit type. When you select a dimension in Looker, only the unique possible values are displayed. As such, the fruit type dimension returns five rows for the 12 fruits, one for each oranges, apples, bananas, lemons and limes. Even though you can see in the basket there are three lemons, two bananas and so on, you don't get multiple rows for them. You only have one row for each. Now, you may also want to know what colors of fruit are in the basket. This would require the color dimension in Looker. Again, only the unique possible values are returned. Instead of getting 12 rows of data, you only get four, one for each for yellow, red, orange and green. Okay, but what if you want to know the combinations of fruit types and colors in the fruit basket? Now, this would require two dimensions in Looker. Notice that you now have six rows. Selecting two or more dimensions returns all unique combinations of the values. For example, the apple type is displayed twice, once for the red apples and again for the green. Be aware that selecting multiple dimensions may result in more rows or different groupings than you may have initially expected. In analytical queries, the dimensions are grouped by the different unique ways in which they can be combined, such as both fruit type and color. A different example might be a reporting of metrics by different cities. In this case, selecting only the city dimension may not give you what you want, as all cities with the same name would group together regardless of their state, province or territory. To separate the cities with the same name, you would need to add the state, province or territory, and you'll then get the individual cities broken out as expected. In summary, dimensions help you to identify and select data attributes that you need to answer your questions in Looker. Each column in a database table is a dimension in Looker, and you can combine dimensions to return all of the unique combinations of the values.
