Data Files

Odoo is greatly data-driven, and a big part of modules definition is thus the definition of the various records it manages: UI (menus and views), security (access rights and record rules), reports and plain data are all defined via records.

Structure

The main way to define data in Odoo is via XML data files: The broad structure of an XML data file is the following:

  • Any number of operation elements within the root element odoo

<?xml version="1.0" encoding="UTF-8"?>
<!-- the root elements of the data file -->
<odoo>
    <operation/>
    ...
</odoo>

Data files are executed sequentially, operations can only refer to the result of operations defined previously

注解

If the content of the data file is expected to be applied only once, you can specify the odoo flag noupdate set to 1. If part of the data in the file is expected to be applied once, you can place this part of the file in a <data noupdate=”1”> domain.

<odoo>
    <data noupdate="1">
        <!-- Only loaded when installing the module (odoo-bin -i module) -->
        <operation/>
    </data>

    <!-- (Re)Loaded at install and update (odoo-bin -i/-u) -->
    <operation/>
</odoo>

Core operations

record

record appropriately defines or updates a database record, it has the following attributes:

model (required)

name of the model to create (or update)

id

the external identifier for this record. It is strongly recommended to provide one

  • for record creation, allows subsequent definitions to either modify or refer to this record

  • for record modification, the record to modify

context

context to use when creating the record

forcecreate

in update mode whether the record should be created if it doesn’t exist

Requires an external id, defaults to True.

field

Each record can be composed of field tags, defining values to set when creating the record. A record with no field will use all default values (creation) or do nothing (update).

A field has a mandatory name attribute, the name of the field to set, and various methods to define the value itself:

Nothing

if no value is provided for the field, an implicit False will be set on the field. Can be used to clear a field, or avoid using a default value for the field.

search

for relational fields, should be a domain on the field’s model.

Will evaluate the domain, search the field’s model using it and set the search’s result as the field’s value. Will only use the first result if the field is a Many2one

ref

if a ref attribute is provided, its value must be a valid external id, which will be looked up and set as the field’s value.

Mostly for Many2one and Reference fields

type

if a type attribute is provided, it is used to interpret and convert the field’s content. The field’s content can be provided through an external file using the file attribute, or through the node’s body.

Available types are:

xml, html

extracts the field’s children as a single document, evaluates any external id specified with the form %(external_id)s. %% can be used to output actual % signs.

file

ensures that the field content is a valid file path in the current model, saves the pair module,path as the field value

char

sets the field content directly as the field’s value without alterations

base64

base64-encodes the field’s content, useful combined with the file attribute to load e.g. image data into attachments

int

converts the field’s content to an integer and sets it as the field’s value

float

converts the field’s content to a float and sets it as the field’s value

list, tuple

should contain any number of value elements with the same properties as field, each element resolves to an item of a generated tuple or list, and the generated collection is set as the field’s value

eval

for cases where the previous methods are unsuitable, the eval attributes simply evaluates whatever Python expression it is provided and sets the result as the field’s value.

The evaluation context contains various modules (time, datetime, timedelta, relativedelta), a function to resolve external identifiers (ref) and the model object for the current field if applicable (obj)

delete

The delete tag can remove any number of records previously defined. It has the following attributes:

model (required)

the model in which a specified record should be deleted

id

the external id of a record to remove

search

a domain to find records of the model to remove

id and search are exclusive

function

The function tag calls a method on a model, with provided parameters. It has two mandatory parameters model and name specifying respectively the model and the name of the method to call.

Parameters can be provided using eval (should evaluate to a sequence of parameters to call the method with) or value elements (see list values).

<odoo>
    <data noupdate="1">
        <record id="partner_1" model="res.partner">
            <field name="name">Odude</field>
        </record>

        <function model="res.partner" name="send_inscription_notice"
            eval="[[ref('partner_1'), ref('partner_2')]]"/>

        <function model="res.users" name="send_vip_inscription_notice">
            <function eval="[[('vip','=',True)]]" model="res.partner" name="search"/>
        </function>
    </data>

    <record id="model_form_view" model="ir.ui.view">
        ...
    </record>
</odoo>

Shortcuts

Because some important structural models of Odoo are complex and involved, data files provide shorter alternatives to defining them using record tags:

template

Creates a QWeb view requiring only the arch section of the view, and allowing a few optional attributes:

id

the view’s external identifier

name, inherit_id, priority

same as the corresponding field on ir.ui.view (nb: inherit_id should be an external identifier)

primary

if set to True and combined with a inherit_id, defines the view as a primary

groups

comma-separated list of group external identifiers

page

if set to "True", the template is a website page (linkable to, deletable)

optional

enabled or disabled, whether the view can be disabled (in the website interface) and its default status. If unset, the view is always enabled.

CSV data files

XML data files are flexible and self-descriptive, but very verbose when creating a number of simple records of the same model in bulk.

For this case, data files can also use csv, this is often the case for access rights:

  • the file name is model_name.csv

  • the first row lists the fields to write, with the special field id for external identifiers (used for creation or update)

  • each row thereafter creates a new record

Here’s the first lines of the data file defining country states res.country.state.csv

"id","country_id:id","name","code"
state_au_1,au,"Australian Capital Territory","ACT"
state_au_2,au,"New South Wales","NSW"
state_au_3,au,"Northern Territory","NT"
state_au_4,au,"Queensland","QLD"
state_au_5,au,"South Australia","SA"
state_au_6,au,"Tasmania","TAS"
state_au_7,au,"Victoria","VIC"
state_au_8,au,"Western Australia","WA"
state_us_1,us,"Alabama","AL"
state_us_2,us,"Alaska","AK"
state_us_3,us,"Arizona","AZ"
state_us_4,us,"Arkansas","AR"
state_us_5,us,"California","CA"
state_us_6,us,"Colorado","CO"

rendered in a more readable format:

id

country_id:id

name

code

state_au_1

au

Australian Capital Territory

ACT

state_au_2

au

New South Wales

NSW

state_au_3

au

Northern Territory

NT

state_au_4

au

Queensland

QLD

state_au_5

au

South Australia

SA

state_au_6

au

Tasmania

TAS

state_au_7

au

Victoria

VIC

state_au_8

au

Western Australia

WA

state_us_1

us

Alabama

AL

state_us_2

us

Alaska

AK

state_us_3

us

Arizona

AZ

state_us_4

us

Arkansas

AR

state_us_5

us

California

CA

state_us_6

us

Colorado

CO

For each row (record):

  • the first column is the external id of the record to create or update

  • the second column is the external id of the country object to link to (country objects must have been defined beforehand)

  • the third column is the name field for res.country.state

  • the fourth column is the code field for res.country.state