[erlang-questions] Failure semantics of dets:insert_new vs dets:insert

[Anthony Ramine]

Looking at the code, I don't see how dets:insert/2 can ever return an error instead of crashing with badarg.

-- 
Anthony Ramine

Le 27 juin 2014 à 16:26, Alexei Sholik <alcosholik@REDACTED> a écrit :

> The signature for dets:insert_new is
> 
>     insert(Name, Objects) -> ok | {error, Reason}
> 
> However, in my use case, I'm using a hash of the value as the key. So I thought I'd switch to using insert_new because that might be slightly more efficient (no need to update stored value).
> 
> But the signature for insert_new is
> 
>     insert_new(Name, Objects) -> boolean()
> 
> In this case, false only indicates that a value for the key is already in the table. What about failure cases of insert()? 
> 
> With insert() I have this code
> 
>    case dets:insert(table, {Key,Value}) of
>       ok  -> ok;
>       {error, _}=Other -> Other
>    end
> 
> With insert_new it turns to just
> 
>    dets:insert_new(table, {Key,Value}),
>    ok.
> 
> IOW, it loses error handling.
> 
> Does insert_new() somehow handle any errors that could arise when calling insert()?
> 
> -- 
> Best regards
> Alexei Sholik
> _______________________________________________
> erlang-questions mailing list
> erlang-questions@REDACTED
> http://erlang.org/mailman/listinfo/erlang-questions