Re: proposal : cross-column stats

2010/12/24 Florian Pflug <fgp@phlo.org>:

> On Dec23, 2010, at 20:39 , Tomas Vondra wrote:
>
>>   I guess we could use the highest possible value (equal to the number
>>   of tuples) - according to wiki you need about 10 bits per element
>>   with 1% error, i.e. about 10MB of memory for each million of
>>   elements.
>
> Drat. I had expected these number to come out quite a bit lower than
> that, at least for a higher error target. But even with 10% false
> positive rate, it's still 4.5MB per 1e6 elements. Still too much to
> assume the filter will always fit into memory, I fear :-(

I have the impression that both of you are forgetting that there are 8
bits in a byte. 10 bits per element = 1.25MB per milion elements.

Nicolas