Use it · 12

A table by dragging, in Palmtree

For the people who live in a pivot table, or once lived in Cognos Query Studio. They know their data and they do not know this tool. So the five things both tools taught them, group, total, pivot, filter and look at the rows behind a number, work the way they expect, and the model does the job a specialist once did by hand in Framework Manager.

1. The field list comes from the model. Each thing with its attributes and each event with its numbers, under the words the business uses. Which fields are numbers is read from the type the model declares, never guessed from the values or from a name.
2. They drag fields in, as they always have. A number goes to Values with the roll-up its type earned: a sum, an average where the model says it is a rate, the ratio of the sums where the model declares a quotient. Anything else groups. The roll-up is on the field and theirs to change.
3. The lines and every total are asked for separately. A subtotal is the same request with fewer groupings and the grand total is the same request with none, so each total is the database's own answer and never a sum of the page.
4. Each request is written from the model and run on the star, read-only. Joins come from what the model declares. The lines come a page at a time; the totals always read every row.
5. The table says what one row means, and its totals are checked. One sentence above the table changes with every drag. Subtotals and the grand total are there without asking, and because they were computed apart from the lines they are held against them: a table whose lines and totals disagree says so. A column that is not a plain sum says what it is, an average or a distinct count, on the column.
6. Double-click a number for the rows behind it. The same request with the adding-up taken off and that cell's own headings pinned, so the rows shown are the rows that were added.

The moves they already know

The move	Query Studio calls it	A pivot table calls it	Here
Put a field in	Insert data	Drag to an area	Drag it to Rows, Columns or Values
Aggregate a column	Summarize	Summarize Values By	The roll-up on the number, set from its declared type
Total and subtotal	A footer; group and summarize	Grand totals; subtotals	On by default, each its own query
Turn a list into a matrix	Pivot	The Columns area	Move the field to Columns; no new query is run
Read a number as a share	Calculate	Show Values As	A reading on the number: share of total, share of parent, difference, running total, rank
The biggest ten	Top or bottom	The Top 10 filter	Top N: a sort and a cut in one move, with the totals still over everything
See the rows behind a figure	Drill through	Show Details	Double-click the number
Two subject areas at once	A multi-fact query	Two tables in one model	Numbers from two stars side by side, as in scenario 3
Keep the report	Save the report	Save the workbook	A saved view

What burned them before, and what happens here

Before	Here
A cross-product error that named nothing, or a silent cartesian product	One sentence in their own words saying what the number can be seen by, and the move that gets them there
An aggregate chosen by a modeler they could not reach, or guessed from the values, so a rate got summed	The declared type decides, it shows on the number, and it is theirs to change
A blank cell that might mean zero, nothing, or a broken relationship	A cell with nothing behind it is drawn as absent and never as a zero
A join path picked by cost or direction and never shown	Where two tables join more than one declared way, the route taken is named and the others are offered