19.1. Converter Basics¶
Converters and SimpleFeatureTypes are defined as HOCON
files. GeoMesa uses the TypeSafe Config library to load the configuration
files. In effect, this means that converters should be defined in a file called application.conf
and placed at
the root of the classpath. In the GeoMesa tools distribution, the files can be placed in the conf
folder. See
Standard Behavior for more information on how
TypeSafe loads files.
19.1.1. Defining SimpleFeatureTypes¶
In GeoTools, a SimpleFeatureType defines the schema for your data. It is similar to defining a SQL database table,
as it consists of strongly-typed, ordered, named attributes (columns). The converter library supports
SimpleFeatureTypes defined in HOCON.
SimpleFeatureTypes should be written as objects under the path geomesa.sfts
.
The name for the SimpleFeatureType will be the name of the HOCON element (e.g. ‘example’, below), or it can
be overridden with type-name
.
A SimpleFeatureType definition consists of an attributes
array, and an optional user-data
section.
attributes
is an array of column definitions, each of which must include a name
and a type
.
See GeoTools Feature Types for supported types. See Reserved Words for names that aren’t supported.
Any additional keys beyond those two will be set as user data, and can be used to configure various
attribute-level options.
The user-data
element consists of key-value pairs that will be set in the user data for the SimpleFeatureType.
This can be used to configure various schema-level options.
See Index Configuration for details on the configuration options available.
Example:
geomesa = {
sfts = {
example = {
type-name = "example"
attributes = [
{ name = "name", type = "String", index = true }
{ name = "age", type = "Integer" }
{ name = "dtg", type = "Date", default = true }
{ name = "geom", type = "Point", default = true, srid = 4326 }
]
user-data = {
option.one = "value"
}
}
}
}
This example is equivalent to the following specification string:
SimpleFeatureTypes.createType("example",
"name:String:index=true,age:Integer,dtg:Date:default=true,*geom:Point:srid=4326;option.one='value'")
19.1.2. Defining Converters¶
A converter defines the mapping between source data (CSV, JSON, XML, etc) and a SimpleFeatureType. The converter
accepts as input source files, and outputs GeoTools SimpleFeatures, which can then be written to GeoMesa.
Thus, each converter corresponds to a single SimpleFeatureType, although there may be multiple converters for each
SimpleFeatureType. The converter library supports converters defined in
HOCON. Converters should be written as objects under
the path geomesa.converters
.
Converters are generally defined with a type
and a fields
array. Optionally,
they may define an id-field
, user-data
and configuration options
.
The type
element specifies the type of the converter, for example ‘delimited-text’ or ‘json’. Specific converters
will have additional options that are not covered here. See GeoMesa Convert for more information on the types available.
The fields
array defines the attributes created by the converter. Each field consists of a name
and
an optional transform
. Specific converters support additional field options; see the documentation
on each converter type for details.
If the name
of a field corresponds with the name
of a SimpleFeatureType attribute, then it will be set as
that attribute when converting to SimpleFeatures. Intermediate fields may be defined in order to build up
complex attributes, and can be referenced by name in other fields.
The transform
of a field can be used to reference other fields or modify the raw value extracted from the
source data. Other fields can be referenced by name using $
notation; for example, $age
references the
field named ‘age’. Transforms can also include function calls. GeoMesa includes a variety of useful transform
functions, and supports loading custom functions from the classpath. See Transformation Function Overview for details.
The id-field
element will set the feature ID for the SimpleFeature. It accepts any values that would normally
be in a field transform
, so it can reference other fields and call transform functions. A common pattern
is to use a hash of the entire input record for the id-field
; that way the feature ID is consistent if the
same data is ingested multiple times. If the id-field
is omitted, GeoMesa will generate random UUIDs for
each feature.
The user-data
element supports arbitrary key-value pairs that will be set in the user data for each SimpleFeature.
For example, it could be used to specify feature-level Accumulo Visibilities.
The options
element supports parsing and validation behavior. See Parsing and Validation for details.