Syntax

   FACTORY_DEF <ReaderKeyword> TilingFactory
   [FACTORY_NAME <factory name>]
   [INPUT FEATURE_TYPE <feature type>
      [<attribute name> <attribute value>]*
      [<feature function>]*]*
   [GROUP_BY [<attribute name>]+]*
   [FLUSH_WHEN_GROUPS_CHANGE (Yes|No)]
   TILE_WIDTH <width>
   TILE_HEIGHT <height>
   [MINIMUM_MEMBERS <members>]
   [ACCUMULATE_ATTRIBUTES (NONE|ALL|ONE)]
   [LIST_NAME <list name>]
   [LIST_ATTRS_TO_INCLUDE [<attrNameToIncludeInList>]* ]
   [LIST_ATTRS_TO_INCLUDE_MODE (ALL|SELECTED)]
   [OUTPUT TILED FEATURE_TYPE <feature type>
      [<attribute name> <attribute value>]*
      [<feature function>]*]* 

Overview

   This factory creates aggregates from its input features, grouping them into 
   tiles covering approximately the width and height specified in the factory 
   definition.

   This factory is useful to reduce the data volume of wallpaper types of 
   features that have no individual attributes. The resulting aggregates, or 
   tiles, can be output to a system using many fewer records than if each 
   feature was output by itself. For systems that support aggregates, or multi-
   part features, this can result in substantial performance improvements and 
   greatly decrease storage requirements.

   The factory uses the tile <width> and <height> which are measured in ground 
   units to divide the input space into cells. The center of the bounding box of 
   each input feature is used to determine the cell for the feature. Once all 
   input features have been read in, an aggregate feature is created from all 
   features in each cell. If linear features are input, they will have pseudo 
   nodes removed within their cells to further reduce the number of separate 
   entities. No such reduction is done to any polygons or donuts that enter. 

   If the MINIMUM_MEMBERS clause is specified, then tiles that have fewer than 
   this number of features plus one (<members> + 1) in them before pseudo-node 
   removal will be merged with a vertically neighboring tile to increase the 
   number of members.

   Features that leave the factory have only the group-by attributes present on 
   them. Any other feature attributes are lost.

   If the GROUP_BY clause is specified, and FLUSH_WHEN_GROUPS_CHANGE is set to Yes,
   the factory will flush the current group whenever a feature from a different
   group is encountered. 

   The ACCUMULATE_ATTRIBUTE clause is responsible for choosing an attribute accumulation
   option. If NONE is selected, the transformer will drop all incoming attributes; if
   ALL is selected, the transformer will merge all incoming attributes and in case of
   conflict, it will preserve attributes of the first incoming feature; if ONE is selected,
   the transformer will get attributes from the first incoming feature.

   The optional LIST_NAME clause is used to associate attributes with incoming feature.
   When the output feature is created, all attributes of
   each feature joining the output are added as members of the specified
   attribute list. The index in the list corresponds to the index of the
   feature's geometry in the aggregate.

   The optional LIST_ATTRS_TO_INCLUDE_MODE will specify if ALL attributes should be added to
   a list, or if only the SELECTED attributes (specified with the LIST_ATTRS_TO_INCLUDE clause)
   should be added to the list.  If this value is not specified, it defaults to ALL.

Assumptions

   The factory does not clip any input features. This means that a feature that 
   spans multiple cells causes one tile to overlap with its neighbors. It is 
   assumed that the input data does not contain features considerably larger 
   than the tile width and height.

Output Tags

   The TilingFactory supports the following output tag.
   
   TILED 
   
      The tiles formed from the input features.