For example, you might want to test various scale factors when displaying multi-value graphs. Let's try this question: Show the discount and product for each order with quantity > 75. When we graph this response, we find it's not shown in an easy-to-comprehend way. For one thing, the Discounts are completely hidden behind the Quantities. We can try to fix this by opening the Chart panel, double-clicking Discount to remove it (from the top of the value list) and clicking the Discount heading to add it back to the bottom of the list. Since we've reversed the order of values, the position of the bar graphs reverses when we click the Graph button. That succeeds in placing the Discounts in front -- but it hardly helps . . . the discount values are too small to see. Of course! That's because discounts are really percentages -- they're always between zero and one. To make the display of the two values more consistent with each other, we can scale the discount by a factor of 100.

Here's where the trick about adding a new query comes in. Rather than modifying the current query, we'll add a new query, and have it inherit all the previous query's values. Butt we'll change the Discount column to read D: Discount * 100. Clear the Chart panes (so that ELF won't go looking for Discount, when it's now called D).When we graph this query, we can see we're making progress. Moving the D value in front of Quantities gives a pretty balanced graph. We can touch it up even more by changing the scale factor to 500. That's easy to do -- just type 500 over the 100 in the SQL pane, and press Enter. Then reclick Graph Selected.