BooleanField Indexed

[Bart Pietercil]

Hi Ruslan,

On 3-okt-07, at 19:26, Ruslan Zasukhin wrote:

> On 3/10/07 5:55 PM, "Ivan Smahin" <ivan_smahin at paradigmasoft.com>  
> wrote:
>
>>> why is a boolean field always indexed when I create it in VStudio.
>>
>>> When I create it in code and I do not specify Indexed it is created
>>> without index (when looking at the table in VStudio)
>>> It seems to me that this is correct. If I don't need an index on a
>>> bool why create the overhead of maintaining an index (as VStudio  
>>> does)?
>>
>>> Why the difference? What am I missing?
>>
>> Of course, it should be no difference.
>> We  don't need index for boolean field at all thanks to special  
>> set of
>> algorithms  for  this  kind  of  field. It would be waste of disk  
>> space
>> only. So even trying to make such field indexed you get no index.
>>
>> Seems to be minor glitch but should be fixed.
>> Please add request to Mantis.
>
> Ivan, let me correct you.
>
> Boolean field in Valentina
>     * ALWAYS is marked indexed,
>     * but DO NOT have index in fact on disk
>
> This is because we keep fields and boolean filed in such way that  
> it is as
> effective as index or even better.
>
>
> This is NOT glitch, and this is NOT bug.
> This is PERFECT UNIQUE feature of Valentina !!! :-)
>
> I have close that report in the Mantis.

But why is the behaviour different when the field is created through  
code (classway).
Would it not be more consistent if a booleanfield when it is created  
in the kernel get the indexflag assigned automaticly ?
When I create it now in RB it does not get the indexedflag unless I  
specify it explicitly.


Bart