TypeConverters

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:

1. String, char, Character

2. All numbers types (primitive + boxed)

3. byte[]/Byte

4. Blob (DBFlow's version)

5. Date/java.sql.Date

6. Booleans

7. Model as @ForeignKey or @ColumnMap

8. Calendar

9. BigDecimal

10. UUID

TypeConverter do not support:

1. Any Parameterized fields.

2. List<T>, Map<T>, etc. Best way to fix this is to create a separate table relationship

3. Conversion from one type-converter to another (i.e JSONObject to Date). The first parameter of TypeConverter is the value of the type as if it was a primitive/boxed type.

4. Conversion from custom type to Model, or Model to a supported type.

5. The custom class must map to a non-complex field such as String, numbers, char/Character or Blob

Define a TypeConverter

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.

TypeConverter for specific @Column

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.