Date Facets

Using virtual date faceting saves on index, while providing greater flexibility.

For example, let us say that each document has a date/time of:

2013/05/12 15:05:12

Before virtual date faceting, date synthesis required that the analysis preprocess all dates and create a category per year, month, and day for them, as well as one per hour, minute, second if time synthesis was required.

Now, with virtual faceting, no specific analysis is required. In your search logic, you can define which categories to create. For example, for year and month faceting, create a DateFacet with withYear and withMonth set to true, and all others set to false.

Synthesis can be enabled independently for year, month, week, day, hour, minute, and second.

• For full dates, refinements are done efficiently with a numerical search.

• For other date-time combinations, refinements are done with a less efficient attrnum query.

When sorting date facets by date, you can control the sort on a category by category basis.

For example, by default a date facet sorts categories by:

• years in descending order

• months in ascending order

• days in ascending order

  
  
  

In the Mashup UI’s Refinement panel, the date facets are displayed like this:

However, you can change this by changing the arrows for each category in the facet. In the following example, years, months, and days are displayed in ascending order.

Use Dynamic Date facets to:

• Drill down in the Refinements panel. Depending on the date granularity you selected for the facet, you can drill from year to month, or from month to days. By contrast, static Date facets display all levels in a hierarchy.

  Dynamic Date Facets Allow You to Drill down in the Refinements Panel
  
  

• Control how many date values display in widgets.

  Widgets have a limited width, which makes it hard to display large date ranges in a legible way. Dynamic date facets solve this by choosing a granularity that displays the most date values possible for the date range, without exceeding the maximum date values allowed. By default, this maximum is 40 but you can change it.

  For example, if the search results included 1-year worth of documents, the facet initially displays 12 months. It uses a month granularity because displaying four quarters does not take full advantage of the number of date values allowed (40 by default), while displaying 52 weeks exceeds it.

  However, if search results only included 2 weeks’ worth of documents, the facet adapts to the shorter period by initially displaying 14 days.

• Fill in missing dates with null values, which is useful for time-series widgets.

  If some documents are missing dates, there is an option to generate missing dates automatically, with a count of 0 and an aggregation value of NO_VALUE.

  Note: You can only generate missing dates when sorting the dynamic date facets by date. The auto-generated dates do not display in the Refinements panel, but they do display in Mashup Builder chart widgets.