Pivoting, Charting
Pre-release
This release adds support for pivoting and charting the results. There is now a third option in addition to the Raw Data and Table views in the Results page: Show as Chart.
This view has two sub-views:
-
Chart - A simple bar chart option that discovers the independent and dependent columns using Chart.js through ember-cli-chart. This option is disabled for Raw queries
-
Pivot - A pivot table/charter using the pivottable plugin that has a lot of C3 based renderers (bar, area, scatter, line) and table renderers (pivoting, table barchart, heat maps etc). It provides a drag and drop interface using jquery-ui. This also enables a bunch of post result aggregations that can applied on to the result data.
The Pivot view is meant to be an advanced use-case. For most Grouping, Top K and Distribution queries, the Chart view should be sufficient. The Pivot view is useful if you have a lot of dimensions and you want to post-aggregate your results, or you want to exclude some outliers for graphing etc.
The result object now stores some additional information that displays what the query was that generated it. This was needed to chart it properly. The UI also stores the pivot table configuration per result. As a result of this, the modelVersion has been bumped to 2 in the env-settings.json properties.
There is a new migrations object that provides a deletions key that can be set to result or query to wipe all results or all data respectively. This is only applied if the modelVersion is not present in localStorage or the modelVersion in settings is greater. Since the result interface has changed, it is necessary to wipe all results as part of migrating to this version.