Before Spotfire version 3.3, time hierarchies were treated just like all other hierarchies. In version 3.3, a new built-in hierarchy is available specific to dates and times.
They use the new BinByDateTime() function to create the expression as shown in the image below.
With the new support for time series hierarchies, visualizations which have a time series on one of their axis will now display a slider by default. See the image below which shows the pre 3.3 visualization on top and then how it looks in 3.3.
When these hierarchies are added to a categorical axis, they act like all other hierarchies. However, some additional settings have been made available to show them more naturally like time series hierarchies, including how to handle gaps in the values, like missing months for example.
To understand what each of the three settings does, we will show 3 visualizations which are identical except for this having a different value for this setting.
These visualizations below show data from July 2005 through June 2007 and have nothing filtered out.
You can see the first two are identical when nothing is filtered out and the last one, using the 'Show all values' option, includes all the months in 2005 and 2007, even though the data does not exist for those months before July 2005 and after June 2007.
In the next set of visualizations, we filtered out all the months in 2006.
The top visualization, with the ‘Show filtered values’ option set, only displays months on the x-axis which are not filtered out and which exist in the data. So, it skips from December 2005 to January 2007 and does not leave any gaps or space between them.
The middle visualization uses the ‘Show filtered range’ option. It will display a gap for all the months in 2006 with no data.
The bottom visualization uses the ‘Show all values’ option. It will display both the gap for all the months in 2006, and it will also show all the months in 2005 and 2007.
In next week’s tip, we will build onto this one and learn how Spotfire handles things like cumulative sums in cases when there are missing time series values.
