can I query raw bits using bit logic in a huge RS database?

[Ruslan Zasukhin]

On 6/10/12 9:33 PM, "Aaron Andrew Hunt" <aaronandrewhunt at gmail.com> wrote:

>> From: Norman Palardy
>> Almost sounds like you need a table that is just 3 columns for the bits like
>> 
>>         master_id
>>         sequence
>>         bit
>> 
>> But queries could get weird as you'd need to write them out in very long form
>> ie/
>>   where sequence = 1 and bit = true
>>      and sequence = 2 and bit = false
>> 
>> etc
>> BUT this could be indexed (which I suspect bit fields really won't be or wont
>> have a useful index)
> 
> 
> Exactly; I came to a similar conclusion but didn't consider having more than 2
> columns. Obviously, I'm just learning how all this works.

I very doubt in this solution guys.

Because 

Then your 3 billions records will blow out to
    3 * 300 avg bits = 900 billions  or more ...

And index on such huge field will be not helpful,
I'd say reverse.


Problem is that you ONE BIT of data
 in the index becomes ( 4 + 4 ) * 1.5 = 12 BYTES .

So index will be crazy huge ...

And even your table ...
    ONE BIT of original data
    now will be bundles with at least 2 bytes  SEQUENCE fields.

This is exactly not a solution


-- 
Best regards,

Ruslan Zasukhin
VP Engineering and New Technology
Paradigma Software, Inc

Valentina - Joining Worlds of Information
http://www.paradigmasoft.com

[I feel the need: the need for speed]