Even more Any and All
I was talking to a few people last week and one of them suggested another change in the any/all syntax. The more I looked at it the more I liked it. In fact when I asked a few of people over twitter it seemed everyone else liked it to.
The suggestion was to replace this:
~/Movies/?$filter=any(Actors a: any(a/Awards))
With this:
~/Movies/?$filter=Actors/any(a: a/Awards/any())
There are a couple of things I think are nice about this:
- It looks a lot like code you might write in something like C# after replace '.'s with '/'s - as is the OData idiom - which hopefully will make it easier to learn.
- It separates the collection variable (i.e. 'Actors') from the range variable (i.e. 'a'). I think this helps because using the old syntax people might be conditioned by the language they use every day to write: any(Actor a: …) instead of any(Actors a: …)
So I think we should do this.
But what do you think?
Alex
As someone that has communicated the need to query over lists with OData, I am very pleased to hear you are working on this. The inability to do this operation is a non-starter for OData in many applications. <br/> <br/> In XPath this query expression would look something like:<br/> <br/> Movies[/Actors/Actor][Awards/Award]<br/> <br/> Personally I am a huge XPath/XSLT fan and have been for many years. If you could match the expressiveness of XPath, at least for descendant operations, with OData, we'd have all the expressiveness we need.<br/>
Looks good and the new syntax reminds me of lambda in a way.<br/> <br/> I give my blessing if that means anything.