The last session I attended on monday was:
The .NET Language Integrated Query (LINQ) Framework
by Luca Bolognese (Senior Program Manager, Microsoft).
For anyone thas has worked in data access components, software, or have struggled at anytime with databases, this represents a huge step forward in terms of productivity, robustness, understandability, and efficiency.
I´m no expert in such issues but, to my eyes, it´s just as if you could write SQL sentences right in the middle of your C# code, combined with the power of .net of course. In addition to that, LINQ can work against almost any kind of data source (SQL data bases, xml, arrays, collections, etc), and serve as well as a datasource for dataBinding.
An example:
[Assume "names" is a collection of strings, an array or whatever]
List<string> query = from s in names
where s.Length == 5
orderby s
select s.ToUpper();
The full LINQ sentence is evaluated as a single expression, but is executed in steps. First, every string in "names" is selected and returned to the second sentence. This one applies a filter selecting those with length == 5 and returns another collection, which is ordered in the third sentence. And so on... as complex as you may need.
I asked Luka how the debugger behaves in this sentences, and he said that LINQ sentences are watched in a special window which allows to see each step of the execution. This is something I´m definately waiting to try.
Just as a last note, LINQ doesn´t always behave this way, as this would be very inefficient. If the collection it´s parsing is just that, something that implements iEnumerable, it behaves so, parsing each element in the collection, iterating through the full array.
BUT, the interesting thing is: if the collection being parsed also implements iQueryable, it behaves differently, building a full query from the LINQ sentences and executing it, as once, at the end. This obviously would make debuggin more difficult, but is the efficient way to execute certain types of queries.
Cheers!
