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
<!-- the root elements of the data file -->
<?xml version="1.0" encoding="UTF-8"?>
<odoo>
<operation/>
...
</odoo>
Data files are executed sequentially, operations can only refer to the result of operations defined previously
Note
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
fieldstype
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 thefile
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 valuechar
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 attachmentsint
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 asfield
, 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 name="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 ainherit_id
, defines the view as a primarygroups
comma-separated list of group external identifiers
page
if set to
"True"
, the template is a website page (linkable to, deletable)optional
enabled
ordisabled
, 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 forres.country.state
the fourth column is the
code
field forres.country.state