indexing gurus pls help me out

[sangoku]

Hy i have  a table which will have GIANT traffic on it like 400q+ in sec i can ease it up with query cashing but it will still have BIG load on it.

The table is made out of 2 int fields and one timestamp.

One stores ip addresses one rand generated numbers. So i am asking what is the best way to index the table... the access is totally random so i need single seek access mode. I was thinking of an hash index on it what do you think?

The table will be always half full and there will be double entry's in both colons because the primary key  are both columns what key block size should i set?

[fenway]

You will have, or you do have?

 

How are you querying this table?

[sangoku]

I will have  :-\ and i will make queries like

 

 

select * from where col1 = some value or something like that

[fenway]

That's not helpful -- what column, and what value?

[sangoku]

sorry to ask... but why do you need to know the exact values???  

 

because i am only asking if i should or should now use a hash index for it and what key size i should use for the hash index..... as for exact values the one is a cookie value and other one is to int converted IP value so

 

the table will have an index which is targeting the values in the ip and cookie value row so what key block size should i set...what is there more you need to know...

[fenway]

Since you don't know the answer, you can't determine which pieces of information are relevant or not.

 

What makes you think you can even make a hash index? Why do you think you need to care about block size?

 

You haven't even explained what this table is for.