Transformer Comparisons

Spatial Transformers

Choosing a Spatial Transformer

Many transformers can assess spatial relationships and perform spatial joins - analyzing topology, merging attributes, and sometimes modifying geometry. Generally, choosing the one that is most specific to the task you need to accomplish will provide the optimal performance results. If there is more than one way to do it (which is frequently the case), time spent on performance testing alternate methods may be worthwhile.

To correctly analyze spatial relationships, all features should be in the same coordinate system. The Reprojector may be useful for reprojecting features within the workspace.

Spatial Transformers Comparison Matrix

Transformer	Can Merge Attributes	Alters Geometry	Counts Related Features	Creates List	Supported Types*	Recommended For
SpatialFilter	Yes	No	No	No	Point Text Curve Area	Testing for the existence of spatial relationships between two sets of features, and routing them according to whether they pass or fail the test(s).
SpatialRelator	Yes	No	Yes	Yes	Point Text Curve Area	Identifying the nature of spatial relationships between two sets of features.
AreaOnAreaOverlayer	Yes	Yes	Yes	Yes	Area	Finding polygon overlaps and extracting them into new geometry.
LineOnAreaOverlayer	Yes	Yes	Yes	Yes	Curve and Area	Finding intersections between lines and polygons, and splitting the lines where they intersect with the polygons.
LineOnLineOverlayer	Yes	Yes	Yes	Yes	Curve	Finding intersections between line features, splitting them, and generating new line geometry as well as points representing the intersections.
PointOnAreaOverlayer	Yes	No	Yes	Yes	Point and Area Text and Area	Identifying points that fall within polygons, and merging attributes between them
PointOnLineOverlayer	Yes	Yes	Yes	Yes	Point and Curve Text and Curve	Identifying where points fall on lines, and splitting the lines into new geometry.
PointOnPointOverlayer	Yes	No	Yes	Yes	Point Text	Identifying points in the same location (within a tolerance), and merging attributes between them.
Intersector	Yes	Yes	Yes	Yes	Point Text Curve Area	Finding intersections between all input features, regardless of geometry (optionally including self-intersections), splitting features, and creating new geometry.
Clipper	Yes	Yes	No	No	Point Text Curve Area Solids Raster Point Cloud	Comparing features against a set of Clipper features, and splitting the features at or along the Clipper boundaries. Outputs both new and untouched geometry, identified as either Inside or Outside the Clipper.
NeighborFinder	Yes	In some cases	No	Yes	Point Text Curve Area	Identifying the nearest other feature(s) to each feature being considered, either in another set of features or within the same feature set.
TopologyBuilder	Yes	Yes	No	Yes	Point Text Curve Area	Analyzing spatial relationships between features to compute topology, splitting features and creating new geometry representing topologically significant nodes, edges, and faces, with associated attributes.

* NOTE: Curve includes Lines, Arcs, and Paths. Area includes Polygons, Donuts, and Ellipses.

Feature Joining

Choosing a Feature Joining Method

Many transformers can perform data joining based on matching attributes, expressions and/or geometry. When choosing one for a specific joining task, considerations include the complexity of the join, data format, indexing, conflict handling, and desired results. Some transformers use SQL syntax, and some access external databases directly. They may or may not support list attribute reading and creation.

Generally, choosing the one that is most specific to the task you need to accomplish will provide the optimal performance results. If there is more than one way to do it (which is frequently the case), time spent on performance testing alternate methods may be worthwhile. Performance may vary greatly depending on the existence of key indexes when reading external tables (as opposed to features already in the workspace).

Joining Transformers Comparison Matrix

Transformer	Match By	Uses SQL Statements	Can Create List	Input Type	Notable	Description
FeatureJoiner	Attributes	No	No	Features	Best performance for simple joins on attribute values Cannot create lists - represents multiple joins as multiple output features Cannot use constructed keys or expressions	Joins features by combining the attributes and/or geometry of features based on common key attribute values. Performs the equivalent of Inner, Left, and Full SQL joins.
FeatureMerger	Attributes	No	Yes	Features	Can use expressions as keys (constructed keys) Can create lists – represents multiple joins as a list on a single feature Cannot match on geometry	Merges the attributes and/or geometry of one set of features onto another set of features, based on matching key attribute values and expressions.
ListBasedFeatureMerger	List Attribute to Single Attribute	No	Yes	Features	Only joining transformer to use all list attribute contents as keys (others may use one specific value in a list, but not the entire list)	Merges the attributes and/or geometry of one set of features onto another set of features, based on matching list attribute values with key attribute values and expressions.
InlineQuerier	SQL query	Yes	No	Features	Use SQL on non-SQL data Allows multiple join/SQL statements, and so can replace multiple transformers with one	Creates a set of SQLite database tables from incoming features, executes SQL queries against them, and outputs the results as features.
SQLCreator	SQL query	Yes	No	External DB	Best if all features already in SQL-capable source Read an external DB table mid-translation	Generates FME features from the results of a SQL query executed once against a database. One FME feature is created for each row of the results of the SQL query.
SQLExecutor	SQL query	Yes	No	External DB	Best for complex joins between features and SQL-capable source Read an external DB table mid-translation	Executes SQL queries against a database. One query is issued to the database for each initiating feature that enters the transformer. Both the initiating features and the results of the query may be output as features.
DatabaseJoiner	Attributes	No	Yes	External DB and Features	Best for simple joins between features and a database table Very efficient if keys indexed, and caches data locally Read an external DB table mid-translation Non-blocking	Joins attributes from an external table to features already in a workspace, based on a common key or keys. SQL knowledge not required. Non-blocking transformer.
Matcher	Geometry and/or Attributes	No	Yes	Features	The only one of these methods that can use geometry as key for matching Can look for attribute differences Uses geometry, not spatial relationships	Detects features that are matches of each other. Features are declared to match when they have matching geometry, matching attribute values, or both. A list of attributes which must differ between the features may also be specified. If matching on attributes only (not geometry), using the FeatureMerger or another method will give better performance.

Raster Transformers

Choosing a Raster Transformer

FME has an extensive selection of transformers for working with raster data. They can be generally categorized as working with whole rasters, bands, cells or palettes, and those designed for workflow control or combing raster with vector data.

Raster Transformers

Working with Rasters

These transformers generally apply to whole rasters.

RasterCellOriginSetter	Sets the raster's cell origin.
RasterConvolver	Accepts input features containing a raster geometry and outputs the features after applying a convolution filter to all the bands. Convolution filters are applied using a kernel, specified by a matrix of weights for the neighboring values, over all the cells of an input raster geometry. Such filtering is used for smoothing, sharpening, enhancing edges and other raster image manipulation operations.
RasterExpressionEvaluator	Evaluates expressions on each cell in a raster, such as algebraic operations or conditional statements.
RasterExtentsCoercer	Replaces the geometry of input raster features with a polygon covering the extents of the raster.
RasterGCPExtractor	Extracts the coordinate system and the Ground Control Points (GCP) from the raster feature and exposes them as attributes.
RasterGCPSetter	Sets the Ground Control Points (GCP) on a raster with the specified Column (pixel), Row (line), X Coordinate, Y Coordinate and Z Coordinate.
RasterGeoreferencer	Georeferences a raster using the specified parameters.
RasterHillshader	Generates a shaded relief effect, useful for visualizing terrain.
RasterInterpretationCoercer	Alters the underlying interpretation of the bands of the raster geometry on the input features, using the specified conversion options. For example, an input raster feature with three bands of interpretation (UInt16, Gray8, and Real64) could be converted to a raster feature with three bands of interpretation (Red8, Green8, and Blue8) or four bands of interpretation (Red16, Green16, Blue16, and Alpha16) in a single operation.
RasterMosaicker	Merges multiple raster features into a single raster feature.
RasterPropertyExtractor	Extracts the geometry properties of a raster feature and exposes them as attributes.
RasterPyramider	Creates a series of pyramid levels for each input raster feature by specifying either the smallest pyramid level size or the number of pyramid levels to generate. Pyramid levels are created by resampling input rasters to various different resolutions.
RasterResampler	Resamples an input raster using the desired dimensions, the desired cell size in ground units, or a percentage of the size.
RasterRotationApplier	Applies the raster rotation angle on the input raster properties to the rest of the raster properties and data values. The expected input is a raster with a non-zero rotation angle and the expected output is a rotated raster with a rotation angle of 0.0. It is expected that the input raster properties will be modified to conform the output raster properties for a raster rotated by the given angle. Applying a rotation angle is primarily done for compatibility with other processing and writers that cannot handle a rotation angle.
RasterSubsetter	Clips raster features using pixel bounds instead of ground coordinates, and optionally adds cells around the perimeter.
RasterTiler	Splits each input raster into a series of tiles by specifying either a tile size in cells/pixels or the number of tiles.
RasterToPolygonCoercer	Creates polygons from input raster features. One polygon is output for each contiguous area of pixels with the same value in the input raster.
WebMapTiler	Creates a series of image tiles that can be utilized by web mapping applications such as Bing™ Maps, Google Maps™, or Web Map Tile Service. This is done by resampling rasters to various different resolutions and then splitting them into tiles.

Working with Bands

These transformers generally apply to bands.

RasterBandAdder	Adds a new band to a raster. The added band will have the same value in all cells, and the same raster-level properties as other bands in the raster (that is, number of rows/columns, cell spacing, cell origin, and so on).
RasterBandCombiner	Merges multiple overlapping raster features into a single raster feature. It accepts a number of input raster features, each of which has one or more bands. The bands are removed from the input features and appended to a single output raster feature.
RasterBandInterpretationCoercer	Alters the underlying interpretation of the selected bands of the raster geometry on the input features, using the specified conversion options. For example, an input raster feature with a single band of interpretation UInt8 could be converted to a single band of Gray8 or UInt16 data.
RasterBandKeeper	Removes all bands of a raster, except for those that are selected. The RasterSelector can be used to modify the selection. If all bands in the source raster are selected, the raster will remain unchanged.
RasterBandMinMaxExtractor	Extracts the band minimum and maximum values, palette minimum and maximum keys and palette minimum and maximum values of a raster feature and exposes them as attributes.
RasterBandNameSetter	Sets the name of selected bands on a raster.
RasterBandNodataRemover	Removes the existing nodata identifier from the selected bands of a raster feature. That is, any values that were previously equal to the nodata value will now be considered valid data.
RasterBandNodataSetter	Identifies a value to act as a nodata identifier on a raster feature at the band level. That is, values equal to the specified value will now be considered invalid, and will not be affected by many operations (for example, offsetting or scaling).
RasterBandOrderer	Specifies the order of bands in a raster. Bands are reordered according to the input band indices.
RasterBandPropertyExtractor	Extracts the band and palette properties of a raster feature and exposes them as attributes.
RasterBandRemover	Removes the selected band(s) of a raster. To modify the selection, see the RasterSelector.
RasterBandSeparator	Separates the bands and palettes from each input raster feature into one or more output raster features based on the number of input bands and palettes.
RasterStatisticsCalculator	Calculates statistics on raster bands and exposes them as attributes. Bands with palettes are valid.

Working with Cells

These transformers generally apply to individual cells.

RasterAspectCalculator	Calculates the aspect (direction of slope) for each cell of a raster. Aspect is measured in degrees from 0 to 360, starting clockwise from the north. Each selected input band will be converted to a Real64 band with output values that represent the aspect. If an input band does not have a nodata value, the output band nodata value will be set to -1.
RasterCellCoercer	Decomposes all input numeric raster features into individual points or polygons. One vector feature is output for each cell in the band.
RasterCellValueCalculator	Performs an arithmetic operation on a pair of rasters.
RasterCellValueReplacer	Replaces a range of values in the source raster with a new single value.
RasterCellValueRounder	Rounds off raster cell values.
RasterSingularCellValueCalculator	Performs an arithmetic operation on two operands: the cell values of a raster and a numeric value.
RasterSlopeCalculator	Calculates the slope (maximum rate of change in z) for each cell of a raster.

Working with Palettes

These transformers generally apply to palettes.

RasterPaletteAdder	Creates a palette from an attribute, and adds this palette to all selected bands on a raster. Selected bands are required to have an interpretation of UINT8, UINT16, or UINT32. Note that palette entries will be discarded if they do not fit within the interpretation of a selected band. For example, when adding a palette to a UINT8 band, all keys that are greater than 255 will be dropped.
RasterPaletteExtractor	Creates a string representation of an existing palette and saves it to an attribute.
RasterPaletteGenerator	Generates a palette out of the selected band(s) of a raster. The output raster will have the selected band(s) replaced by a new band with a palette.
RasterPaletteInterpretationCoercer	Alters the underlying interpretation of the palettes of the raster geometry on the input features, using the specified conversion options. For example, an input raster feature with a single band with a single palette of interpretation RGB24 could be converted to a single band with a single palette of RGB64 or String data.
RasterPaletteNodataSetter	Identifies the nodata value on a raster feature at the palette level. The transformer will succeed in setting the specified nodata value only if the input raster band(s) have at least one palette and a nodata key has already been set on the band.
RasterPaletteRemover	Removes the selected palette(s) of a raster. The palette's band retains the palette key, but the corresponding values are lost.
RasterPaletteResolver	Resolves the palettes of the selected bands of the input raster features by using the band cell values to look up the corresponding palette values, which then replace the original band cell values in the raster.

Workflow Control

These transformers generally control the flow of features in a workspace.

RasterCheckpointer	Sets a checkpoint in the raster processing which forces previous processing to occur immediately. Once complete, it saves the current state to disk.
RasterConsumer	Requests the tile(s) from the raster geometry but no actual operations are performed on the tile(s).
RasterExtractor	Serializes the geometry of the feature into the Blob Attribute based on the selected writer format.
RasterNumericCreator	Creates a feature with a raster of the specified size with a numeric value and sends it into the workspace for processing. It is useful for creating a very large image with a user-specified width and height.
RasterReplacer	Replaces the geometry of the feature with the geometry held in the Blob Attribute. The blob is decoded according to the selected raster format.
RasterRGBCreator	Creates a feature with a raster of the specified size with an RGB value and sends it into the workspace for processing.
RasterSelector	Selects specific bands and palettes of a raster for subsequent transformer operations. The bands and palettes are selected using the band and palette indices, specified in a string. The string may either be specified explicitly or through an attribute. The format of the string is B P (separated with a space), where B is the band index and P is the palette index of the band and palette to be operated on. Indices are zero-based, so the first band or palette is at index 0.

Vectors and Rasters

These transformers generally involve using rasters and vector data together.

ImageRasterizer	Draws input point, line and polygon features onto a color raster filled with the background color. The fme_color attribute of the input vector features is used to generate pixel values. Features without an fme_color attribute will be discarded.
NumericRasterizer	Draws input point, line and polygon features onto a numeric raster filled with the background value. The Z coordinates of the input vector features are used to generate pixel values. Features without Z coordinates will be discarded.
MapnikRasterizer	Draws input point, line, polygon, and raster features onto a raster using the Mapnik toolkit.
PointOnRasterValueExtractor	Extracts the band and palette values from a raster at the location of each input point and sets them as attributes on the feature.
VectorOnRasterOverlayer	Overlays vector features onto a single raster feature by drawing them onto the resulting output raster. The properties of the output raster are identical to that of the input raster.

asdf