Skip to content

Latest commit

 

History

History
74 lines (56 loc) · 3.23 KB

field-converter.md

File metadata and controls

74 lines (56 loc) · 3.23 KB

Field converter

A field converter can alter a field's database schema and convert the related data to a new format. For example in Django it is not possible to alter a ManyToManyField to a CharField or to convert from and to a ManyToManyField. The alter_field function of the schema editor would not work in this case. If you want to convert it you have to create a new column and remove the old one instead of altering. This is something that you do with a field converter.

How it works

When the field type changes or a field's property changes Baserow will check if there is an applicable converter. It does so by looping over the registered field converters, calling the is_applicable method which determines based on the from and to field instances if the converter can be applied in the situation. If an applicable converter is found it will be used. It could be that there are multiple converters applicable, but it will use the first one that can be applied. If there are not any converters applicable then the regular lenient schema editor's alter_field method will be used.

An example

In the example below we are going to check if the from_field is a text field and if the to_field is a date field. If that is the case we first want to remove the old field and then create the new field instead of using the lenient schema editor's alter_field method. Of course this would not make sense in real life, but it is just to demonstrate how a converter could work. You can of course be really creative in the alter_field method of your converter. You can for example first load all the old values in memory before removing the field and later update them again after the new field has been added. You can also first create the new field, perform some query that updates the data of the field based on the old field and then delete the old field and rename the new field. A lot is possible. We do recommend to keep performance in mind. Ask yourself the question does it stay fast if there are 100k rows?

plugins/my_baserow_plugin/backend/src/my_baserow_plugin/field_converters.py

from baserow.contrib.database.fields.field_types import (
    TextFieldType, DateFieldType
)
from baserow.contrib.database.fields.registries import FieldConverter


class TextToDateFieldConverter(FieldConverter):
    type = 'text-to-date'

    def is_applicable(self, from_model, from_field, to_field):
        return (
                isinstance(from_field, TextFieldType) and
                isinstance(to_field, DateFieldType)
        )

    def alter_field(self, from_field, to_field, from_model, to_model,
                    from_model_field, to_model_field, user, connection):
        with safe_django_schema_editor() as schema_editor:
            schema_editor.remove_field(from_model, from_model_field)
            schema_editor.add_field(to_model, to_model_field)

plugins/my_baserow_plugin/backend/src/my_baserow_plugin/config.py

from django.apps import AppConfig

from baserow.contrib.database.fields.registries import field_converter_registry


class PluginNameConfig(AppConfig):
    name = 'my_baserow_plugin'

    def ready(self):
        from .field_converters import TextToDateFieldConverter

        field_converter_registry.register(TextToDateFieldConverter())