Re: How does a table scan effect the cache?

Hello!

Chuck Hamilton <chuck_hamilton_at_yahoo.com> ñîîáùèë â íîâîñòÿõ ñëåäóþùåå:l3gtlt0ivev2n9cml486d0ab99lku7c4j3_at_4ax.com...
> When a full scan is done on a table, it's my understanding that the
> blocks read go to the least recently used end of the LRU list. Does
> that mean that every time the scan reads a group of blocks
> (db_file_multiblock_read_count to be exact), that the same blocks at
> the end of the LRU list keep getting replaced on each read? So
> assuming my dbfmbrc = 16, and no other access is happening against the
> table, and the table is large enough to not be cached at the top of
> the list, I would never have more than 16 blocks of that table in the
> cache? Is my understanding on this correct?

No, It's not;

For example:
You data cache 10M
select /*+ ALL_ROWS */* from YOU_TABLE (need 15M)

at the end of select all 10M will contain last 10M of "select".

i.e. first 5M will be rewrited.

Yaroslav.

> --
> Chuck Hamilton
> chuck_hamilton_at_yahoo.com
>
> "Do not be deceived, God is not mocked;
> for whatever a man sows, this he will also
> reap." (Gal 6:7 NASB)
Received on Wed Jul 25 2001 - 08:55:15 CDT