Hi and thank you for your answer

> That might be the case for Intersystems Caché. I don't think it would help a 
> lot for H2.

True, experience with and behaviour of another DBMS cannot be simply
translated to H2 right away.

Nevertheless, the main advantage of BITMAP indices is of mathematical
nature, not just an implementation detail. That's why other DBMS have
BITMAP indices, too.

An example:
Say you have a query with several column equality checks in the search
condition (e.g year = 2011 AND month = 12 AND lastname = 'Mueller'
AND ... )
With normal indices, each one of them would have to be checked one by
one - or even not at all if the data / selectivity is not suitable.
Having BITMAP indices for those columns, the end result can be
calculated very quickly by AND-ing together the corresponding bitmaps
of ALL the columns and then simply collecting the rows from the
resulting row numbers.
Escpecially for data ranges with low selectivity, but even for high
selectivity (where BITMAP indices are not well-suited) the synergy
effects of the efficient AND-ing can outperform normal indices that
would be better suited for a single column.

I don't want to answer you back (too much :D), but I could imagine
that such math of data selectivity, boolean arithmetic, etc. apply to
H2 as well and could very much improve it.
It's a little like convicing people that a HashSet.contains() really
mathematically is O(1) while a ArrayList.contains() is O(n) :-D.

> Currently not. But if you want to implement it, please go ahead.

Thanks for the honest answer. I actually might give it a try (I
already tinkered around a little on a java bitmap-collection), if I
find time between work and three private library projects :-/.

Regards




On Dec 15, 9:13 pm, Thomas Mueller <[email protected]>
wrote:
> Hi,
>
> > Intersystems Caché:
> > (btw. we found that BITMAP indices are superior to "normal" indices
> > for most data that has to be indexed at all and later on, one of
> > intersys' tech consultants approved that from their point of view)
>
> That might be the case for Intersystems Caché. I don't think it would
> help a lot for H2.
>
> > So all I want to ask now is: Is there an interest of putting the topic on 
> > the roadmap for a future version of H2?
>
> Currently not. But if you want to implement it, please go ahead.
>
> Regards,
> Thomas

-- 
You received this message because you are subscribed to the Google Groups "H2 
Database" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/h2-database?hl=en.

Reply via email to