I tried to test evolution graph with "period=range". For example from 8th August 2012 to 2nd September 2012. It renders a graph displaying "No data for this graph" if there are no visits/actions on the last day of the period. If I set range from 8th August 2012 to 3rd September 2012, it dispalys the graph. There are actions and visits in August. It seems just to depend whether there are actions/visits on the last day.
When the graph is displayed with message "No data for this graph", the graph is also loading fast. Maybe it just looks at the last day and does no further queries in this case.
Manage this edge case when there is no data. This happens with any period, not only range.
In 930ec65093018c13c39091e991a4ebacd7a30ee3: Refs #3465, use bulk tracking in RowEvolution integration test to speed it up.
In 80ac6ac1666508cb0844dc1eac2faba303fa9f9e: Refs #3465, remove Piwik_DataTable_Array::metadata and move data to Piwik_DataTable::metadata.
In 900eec1ff8bfa8ef758df53bb53908fad684448c: Refs #3465, fixing build
In 51b77878acf71118e9d4cb769d72ded7bbfbf6fa: Fixes #3465, refactor row evolution code to fix bug where if no labels are specified and last period has no labels, no data is returned. Refactoring also contains optimization for case where no labels are specified.
In d9ff5001ecfe0fc763e62dc3a9363fc4a45d02d3: Refs #3465, fix regression in multi row evolution API output where columns w/ value of 0 were not outputted.
In 7a89c65378fd7893d5c0b0fdec925b870adb4ed9: Refs #3465, make sure labels are associated with correct rows in multi-row evolution after generic filters (ie, Sort) are applied.
In 204cbc0af2cc2914235f6a720e4a5abf0d6b4101: Refs #3465, fix regression for flattener filter where queued filters were applied on the flattened datatable.
In a8420c06e0e11288a344bfc054c7fb6970db3c7f: Refs #3465, fixing build.