Community
Participate
Working Groups
BIRT reports appear to include two kinds of data sets. There is the explicit one created in the data explorer. The data set editor provides a very nice preview of the data set results which is wonderful for verifying that the data set works as expected. But, BIRT also seems to define a "virtual" data set (VDS) on a table: * The VDS has its own column bindings (which must, for the most part, mimic that of the DS bound to the table, but with changes to data type and display name.) * The VDS can add per-row computed columns in addition to those in the DS. * The VDS adds aggregates. * The VDS adds column bindings for each data item in the report. * The VDS provides implicit grouping (as defined by group levels in the table.) All of this adds up to a healthy chunk of additional functionality that the user must get right. However, the only way to test this VDS is to run the report. Doing so adds an additional level of complexity: the user sees the VDS as presented by the report. If something is wrong, the user must figure out whether the VDS or presentation is at fault. It is this ability to clearly separate presentation and data access that makes the DS preview feature so useful. But, that benefit gets lost somewhat because the additional functionality on the VDS cannot be previewed. Further, BIRT allows aggregates. A newbie user might expect these to be available in the corresponding group and/or table footer. Are they available elsewhere? One can figure this out by experimentation. (Add temporary "dummy" fields into various points in the table to see what value the aggregate has.) Or, one can try to read the documentation. It would be much easier, however, simply to run a preview and look to see what value each aggregate has at various parts in the data set. Does the grand total have a value on the first row? On the 10th row? A picture (of the result set) is worth a 1000 words of on-line help or documentation. Related to this is the question of what values are available in the table header/footer, and the group header/footer. Is the first data row available in the table header? The last row in the table footer? The first row of the group in the group header? And so on. A visualization would answer these questions, especially if the visualization provided some kind of highlighting or labeling to say when each row is active. For example, the first row of the data set might have a label that says "H,GH1,GH2,D" to show that that row is available to the Header, Group 1 Header, Group 2 Header and Detail band. (The actual UI would, one would hope, be more clever than just a text label; but even the text label is better than guessing.) Note that, even if the VDS were defined implicitly (as requested by other bugs), the preview would still be helpful. The VDS is available to the report at runtime regardless of whether the software "compiles" it at runtime, or the user builds it explicitly at design time (as is the case today.)
Created attachment 94153 [details] Product Sales by Country.rptdesign
It turns out that UI does exist to visualize the "virtual data set" on a table: it is in the Chart UI: 1. Create a chart in the header row for a table. 2. Bind the chart to "Inherit from container." 3. Voila! The result set appears in the Data Preview area. Of course, this is not the most convenient general solution: 1. I must add a chart to do the visualization. 2. The header "chicklets" in the table's data preview don't show the full name of the column. 3. The window itself is very small; it shows a small subset of rows and columns. Still, this is a crude-but-possible workaround. And, it shows that creating a proper UI is "simply" a matter of UI; the internal support for it must already exist to power the data preview in the chart UI. Thus, the request stands: provide a simple UI to visualize the table's data without having to add a dummy chart.
+1 for this enhancement. add the plan key word.
Need some UI design.
Moved to future.
