How reporting works#

Early release. This feature is in early release, so its screens and behavior may change. Tell us what you think with Send feedback. See Early release features and Send feedback.

Reportworq reporting is a chain of four stages. Each stage has one job, and each hands something finished to the next. If you keep the chain in mind, every other page in this guide has an obvious place.

The chain on one page#

  1. DATA SOURCES        a planning cube, a SQL database, a workbook
          |
          v
  2. DATA MODEL          one or more sources shaped into named VIEWS
          |
          v
  3. REPORTS             built on views, placed on a web report
          |              or in an Excel workbook
          |
          +-----------------------------+
          |                             |
          v                             v
  4a. VIEWED LIVE                4b. SENT BY A DISTRIBUTION JOB
      by an author or                refreshed, personalized and
      administrator                  delivered as static output
                                     (PDF, Excel, email, and so on)
                                           |
                                           v
                                     READERS receive it

Stage by stage#

1. Data sources#

A source is a connection to a system of record, such as a planning cube or a SQL database, or a file that lives in your workspace. You connect each source once and every model can then use it. Nothing about a source changes when you build a model on it.

2. The data model#

A data model reads one or more sources and turns them into views. A view is one query plus a curated set of fields, and each field has a stable name that survives changes in the source. Several sources can feed one model, and views can be joined across sources. The model is also where you decide which views report authors can see.

Start with Data models, then go to the page for what you need:

• Bring several sources into one model
• Build a view, Joined views and Table relationships
• Hierarchies and Field filtering
• Direct queries and View visibility
• Data model variables and Global parameters
• Discover your data and Ask AI on a data model
• Build a data model from a workspace file

3. Building reports#

A report reads a view and shows it as a list, a crosstab, a chart or a map. Reports are placed on a web report, an Excel-like page you edit in the browser, or in an Excel workbook that you build with the Excel add-in. The two meet in the same studio. New > New Web Report starts a web report.

Start with The web report studio and The Excel add-in and the studio, then:

• List reports, Crosstabs and Charts and maps
• Professional reports
• AI insights
• Combine report types
• Cell-driven variables

4a. Viewing live reports#

Opening a web report in Viewing mode shows it with current data. A person viewing can change what the report shows (filters, expanding and collapsing, drilling to source rows) but not how it is built, and nothing they change is saved. From the viewer you can download the report or split it into parts by a filter. See The viewer and Download, burst and save as.

4b. Sending with a distribution job#

A distribution job is how a report reaches the people who need it. The job refreshes the report with current data, personalizes it for each recipient if you ask it to, and delivers the result by email, a shared folder or another destination. What recipients get is static: a file that shows the numbers as of the run.

A web report reaches a job in one of two ways. The route to prefer is to convert it to an Excel workbook, which the job then refreshes in place on every run. See A web report as a job source, Convert to Excel and Key cell formats and what a job keeps. For the job itself, see the distribution job lifecycle and the distribution overview.

Live and static#

Readers never see web reports. Web reports are working tools for the people who build and analyze. A reader receives the static output of a job. This is by design, so a reader never has to learn the studio and never sees a half-finished report.

Two kinds of splitting exist, and they do not mix. The viewer's Burst pane is a lightweight way for an analyst to cut one live report into parts for themselves. A distribution job is the author's full bursting, and it produces static content for the people on a burst set. A job does not carry the live report's filters. To learn the job side, see What is bursting.

Who can do what#

What needs a license#

An administrator can see which features the license grants on Settings > Configuration > License. See Licensing.

Going deeper. Some of these features are in early release. See Early release features and Send feedback. If something does not behave as described, see Troubleshooting and limits.