Localization

Basic concept
You can define per field if the field is localizable or not. We call this system partitioning, but more about that later.
It is easy to understand when you have a look to an content object from the API:
1
{ 
2
    "id": "01",
3
    "created": "2017-02-25T19:56:35Z",
4
    "createdBy": "...",
5
    "lastModified": "2017-02-25T19:56:35Z",
6
    "lastModifiedBy": "...",
7
    "data": {
8
        "name": {
9
            "en": "Copenhagen",
10
            "sv": "Köpenhamn",
11
            "fi": "Kööpenhamina",
12
            "it": undefined // Not part of the API response
13
        },
14
        "population": {
15
            "iv": 1400000
16
        }
17
    }
18
}
Copied!
Each field value is a set of values that are associated to keys. In JavaScript it is called a object, other programming languages call it HashMap or Dictionary. The keys must be unique. Depending whether the field is localizable or not the API accepts different keys.
The population field is not localizable. Therefore the only allowed key is iv, which stands for "invariant".
The name field is localizable. The allowed keys are the language codes for the languages you have configured.
The languages an fallback rules can be configured in the Management UI:
Settings
In this example we have 4 languages:
Name
Code
Fallback
Description
English
en
-
Our master language. Whenever a fields is not available in a language it falls back to the master language.
Swedish
sv
en
Swedish has no fallback language configured, therefore the fallback language is always English.
Finnish
fi
sv,en
Finish has Swedish configured as a fallback language. This means that whenever a value for a localizable fields is not available Squidex tries to resolve the value from Swedish first and then from the master language (English). Finish is also marked as optional, which mean that required fields can be omitted. This is useful when you introduce a new language. You can save contents with required fields even if the field value has not been entered for the optional language.
Italian
it
en
Italian is a new language and has been added after the content editors have created most of the content. Therefore most content items do not have a value available.
How to use the API
How to retrieve the correct languages?
The rest endpoint provides two headers that can be used to query the correct language.
X-Languages Header
You can filter the languages with the X-Languages header. Other languages will be omitted. We do not use the Accept-Language header because we want to avoid compatibility issues. If you define a language that is not supported, this language will be ignored. For example: If you set: X-Languages: en,sv,de for our example above you will only retrieve English (en) and Swedish (sv).
1
X-Languages: en,sv,it
2
{ 
3
    ...,
4
    "data": {
5
        "name": {
6
            "en": "Copenhagen",
7
            "it": "Copenhagen"
8
        },
9
        "population": {
10
            "iv": 1400000
11
        }
12
    }
13
}
Copied!
If none of the specified languages is provided you will retrieve the master language only.
1
X-Languages: de
2
{ 
3
    ...,
4
    "data": {
5
        "name": {
6
            "en": "Copenhagen"
7
        },
8
        "population": {
9
            "iv": 1400000
10
        }
11
    }
12
}
Copied!
X-Flatten Header
If you add this header, fields that only have a single value will be flattened. So the example above will be transformed to:
1
X-Flatten: true
2
{ 
3
    ...,
4
    "data": {
5
        "name": {
6
            "de": "Copenhagen",
7
            "sv": "Köpenhamn"
8
        },
9
        "population": 1400000
10
    }
11
}
Copied!
Both headers can be combined. If you define a single language with the X-Languages header the localizable fields will contain only one value each and therefore they can be flattened as well. If you provide an unsupported language you will just get the master language.
So our example from above might look like:
1
X-Languages: de, X-Flatten: true
2
{ 
3
    ...,
4
    "data": {
5
        "name": "Copenhagen",
6
        "population": 1400000
7
    }
8
}
Copied!
It basically means that you can just forward the user language and Squidex will handle the rest.
The headers above are not supported by the GraphQL endpoint, because in graphql the output should be defined the query only.
How to disable fallback languages?
If you want to do the fallback handling in your API, you can disable the behavior above. Add the following header to all your to all your requests: X-NoResolveLanguages=1
You will get the raw data then and some fields might not have a value for a language, for example when this language is new and nobody has entered a value yet:
1
X-NoResolveLanguages: 1
2
{ 
3
    ...,
4
    "data": {
5
        "name": {
6
            "en": "Copenhagen",
7
            "sv": "Köpenhamn",
8
            "fi": "Kööpenhamina",
9
        },
10
        "population": {
11
            "iv": 1400000
12
        }
13
    }
14
}
Copied!
Why do you call it partitioning?
It basically means that the a value is partitioned into multiple subvalues. When we implemented the localization feature we realized that it might be very helpful to extend this feature to other type of keys, for example you could...
... define your prices for different currencies.
... write your texts for different countries.
... define customer groups.
So we implemented the localization feature with the idea in mind that we might extend it in coming versions.
Schemas
Migrations
Last modified 2mo ago
Copy link
Contents
Basic concept
How to use the API
How to retrieve the correct languages?
How to disable fallback languages?
Why do you call it partitioning?