This is an old question, and I'd like to know if there is any enhancement to the status quo.
The question is easy: I must build an application which is used by people from different countries, so I have to use different languages.
The static part of the site is solved using i18n.
But when I run a Query against the DB, I receive the result in the language used to create data..
So, for instance, If I have a node with a property name, I can associate just one string to that property, for instance name: "house" or name: "dog".
The luxury top should be to have the opportunity to store tuples, like name: {'en': "dog", 'it': "cane"} then have the ability to specify the language as a query parameter match (n:Object {_lang: 'it'}) return n.name and have cane as the result (or pass/use the lang paramether in the Query header).
Another approach would be to append the language attribute to the property name, like 'name:it: "cane", name:en: "dog"', maybe the latter is easier to manage but a little unconfortable to implement due to the : already used as a separator.
With both the solutions, of course, when an index is built on the property, all the language indexes should be created.
With this approach there should be no reason to introduce dirty structures in the DB design.
Is there any way to reach this kind of result, or, at least, a result that doesn't require to mess a clean DB design and also doesn't introduce an important level of complexity to the Cypher queries?
Of course, any suggestion is welcome, both for me, and both for people working on Cypher maintainers.
Thanks