Am 21.01.10 21:54, schrieb Andrus Adamchik:
>
> On Jan 21, 2010, at 10:38 PM, Andrey Razumovsky wrote:
>
>> So, following specification, this cannot be performed anyhow?
>
> No. But the idea I guess is that pattern is not another column value. It
> is always a literal.
Yes, the tutorial emphasizes this: "The pattern value is a string
literal that may contain wildcard characters."
http://java.sun.com/j2ee/tutorial/1_3-fcs/doc/EJBQL5.html
I find this restriction rather strange, but maybe the authors couldn't
think of a reasonable real-world scenario which requires a dynamically
generated pattern.
-- Andreas
>
>>
>> I agree subselects in LIKE is not good, then maybe let's change that to
>> input_parameter() | string_literal() | functions_returning_strings() ?
>> Specification does not say what to do if right part is not input
>> parameter
>> or string literal. So is it really bad if we do more than specification
>> says?
>
> I think it doesn't buy us much. LIKE is a rather special case IMO, and
> this is reflected in the spec.
>
>> Another weird thing is that pattern_value() unlike other expressions
>> is not
>> described in BNF on pages 109-112..
>
> True. It is described in the text only. The EJBQL BNF is rather sketchy
> in some parts.
>
> Andrus
>
>
-- Andreas Hartmann, CTO BeCompany GmbH http://www.becompany.ch Tel.: +41 (0) 43 818 57 01
This archive was generated by hypermail 2.0.0 : Thu Jan 21 2010 - 16:35:38 EST