(c#)How to implement a high performance duplicate check?

i want to find out a high performance duplicate check way. does anybody can help me ? i implement the check as the below, but now, i hope the speed can catch up sending message into a queue(MSMQ), so the old way is not fit with us .

my old way:

1. create a simple table(oracle db)
2. create PK on the column which i want to do the duplicate checking.

firstly, i will insert a record with try-catch into the table.

if(insert succeed)
{
not duplicate
}
else
{
duplicate
}

PS: the messages need to check is very large, so i can't record them in开发者_Python百科to the memory.

You could try something like managedesent

Check out the persistent dictionary implementation.