When building out Model classes, you may wish to provide a different type of @Column that from the standard supported column types. To recap the standard column types include:
-
String,char,Character -
All numbers types (primitive + boxed)
-
byte[]/Byte -
Blob(DBFlow's version) -
Date/java.sql.Date -
Booleans
-
Modelas@ForeignKeyor@ColumnMap -
Calendar -
BigDecimal -
UUID
TypeConverter do not support:
-
Any Parameterized fields.
-
List<T>,Map<T>, etc. Best way to fix this is to create a separate table relationship -
Conversion from one type-converter to another (i.e
JSONObjecttoDate). The first parameter ofTypeConverteris the value of the type as if it was a primitive/boxed type. -
Conversion from custom type to
Model, orModelto a supported type. -
The custom class must map to a non-complex field such as
String, numbers,char/CharacterorBlob
Defining a TypeConverter is quick and easy.
This example creates a TypeConverter for a field that is JSONObject and converts it to a String representation:
@com.dbflow5.annotation.TypeConverter
class JSONConverter : TypeConverter<String, JSONObject>() {
override fun getDBValue(model: JSONObject?): String? = model?.toString()
override fun getModelValue(data: String?): JSONObject? =
try {
JSONObject(data)
} catch (JSONException e) {
// you should consider logging or throwing exception.
null
}
}
}Once this is defined, by using the annotation @TypeConverter, it is registered automatically across all databases.
There are cases where you wish to provide multiple TypeConverter for same kind of field (i.e. Date with different date formats stored in a DB). You can override a field's TypeConverter locally at the @Column level.
In DBFlow, specifying a TypeConverter for a @Column is as easy as @Column(typeConverter = JSONConverter::class). What it will do is create the converter once for use only when that column is used.