That make sense if it has technical reason :) I thought there's no technical reason, but just because the team want to keep the behavior same as EF. 2014-3-31 下午11:40于 "Diego Mijelshon" <[email protected]>写道：
> A SUM with no rows returns null in SQL. > You are trying to cast NULL to a non-nullable type. That's what causes the > exception. > Automagically transforming null to 0 would be a bug. > Yes, this is not what happens with LINQ to objects. But LINQ is a leaky > abstraction. > > Diego > > > On Mon, Mar 31, 2014 at 11:25 AM, Mouhong Lin <[email protected]>wrote: > >> but make it returns 0 and not throwing exception is not "surprise" right? >> but the exception is "surprise". If in EF user use the same workaround, >> when he turn to NH, that workaroud is still working, so it doesn't break >> anything, so there's no "surprise". >> 2014-3-31 下午9:56于 "Stephen Bohlen" <[email protected]>写道： >> >>> I don't think its technically 'broken'. As you can see from the >>> comments in the JIRA issue, the final design decision was to mimic the >>> behavior of sim. LINQ statements in other ORMs in the .NET space for the >>> sake of consistency and so as not to violate "the principle of >>> least-surprise". >>> >>> If that behavior isn't what you seek for your use-case, then you are >>> certainly free to apply the work-arounds discussed. If you don't like the >>> "ugly" syntax this produces in your code, then I recommend you create an >>> extension method of your own ( perhaps "IEnumerable.NullSafeSum(...)" ?) >>> and invoke that instead. >>> >>> Cheers, >>> >>> Steve Bohlen >>> [email protected] >>> http://blog.unhandled-exceptions.com >>> http://twitter.com/sbohlen >>> >>> >>> On Mon, Mar 31, 2014 at 1:03 AM, Mouhong Lin <[email protected]>wrote: >>> >>>> I already know that workaround before I created this post. >>>> I mean why this is treat as "by design" behavior and not plan to fix it. >>>> That's why I posted it here instead of the nhusers groupd. >>>> >>>> >>>> On Monday, March 31, 2014 4:14:16 AM UTC+8, Ricardo Peres wrote: >>>>> >>>>> var total = query.Sum(x => (double?) x.Income); >>>>> >>>>> Replace double for single, int, long, etc. >>>>> >>>>> By the way, the proper place to ask questions is the nhusers mailing >>>>> list, not this one. >>>>> >>>>> RP >>>>> >>>>> On Sunday, March 30, 2014 4:42:24 PM UTC+1, Mouhong Lin wrote: >>>>>> >>>>>> Hi guys, >>>>>> >>>>>> I get an exception saying "Value cannot be null" when i execute this >>>>>> query (NH 3.3.3): >>>>>> >>>>>> var total = query.Sum(x => x.Income); >>>>>> >>>>>> The exception happens when the query returns no result. >>>>>> I found there's already a Jira issue about this: >>>>>> https://nhibernate.jira.com/browse/NH-3113 >>>>>> >>>>>> But why won't fix? It's really not usable if this query throws >>>>>> exception. In every application that need to do reporting will encounter >>>>>> this. >>>>>> And this violate the semantic of the sum and linq to objects, right? >>>>>> If this should throw exception, why query.Count() don't throws? >>>>>> (I understand the technical reason of this, just saying why the >>>>>> semantic of these are not consistent) >>>>>> >>>>>> If L2S or EF behaves the same, I think it's their bug. Not because >>>>>> it's a good thing. >>>>>> Workaounds like this "var sum = myQuery.Sum(x => (int?)x.Prop) ?? 0" make >>>>>> my code really ugly. >>>>>> >>>>>> Ideas? >>>>>> >>>>>> >>>>>> >>>>>> ------------------------------ >>>>>> Mouhong >>>>>> >>>>>> -- >>>> >>>> --- >>>> You received this message because you are subscribed to the Google >>>> Groups "nhibernate-development" group. >>>> To unsubscribe from this group and stop receiving emails from it, send >>>> an email to [email protected]. >>>> For more options, visit https://groups.google.com/d/optout. >>>> >>> >>> -- >>> >>> --- >>> You received this message because you are subscribed to a topic in the >>> Google Groups "nhibernate-development" group. >>> To unsubscribe from this topic, visit >>> https://groups.google.com/d/topic/nhibernate-development/o5orsYz7Ums/unsubscribe >>> . >>> To unsubscribe from this group and all its topics, send an email to >>> [email protected]. >>> >>> For more options, visit https://groups.google.com/d/optout. >>> >> -- >> >> --- >> You received this message because you are subscribed to the Google Groups >> "nhibernate-development" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> For more options, visit https://groups.google.com/d/optout. >> > > -- > > --- > You received this message because you are subscribed to a topic in the > Google Groups "nhibernate-development" group. > To unsubscribe from this topic, visit > https://groups.google.com/d/topic/nhibernate-development/o5orsYz7Ums/unsubscribe > . > To unsubscribe from this group and all its topics, send an email to > [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- --- You received this message because you are subscribed to the Google Groups "nhibernate-development" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
