Which Oracle access method is the fastest way for Oracle to retrieve a single row?
-
Primary key access
-
Access via unique index
-
Table access by ROWID
-
Full table scan
ROWID access is the fastest method because Oracle uses the row's physical disk location directly, eliminating the need to traverse index structures or scan multiple rows. Primary key and unique index access require additional index lookups, while full table scans read every row.
Table access by ROWID is correct. A ROWID is Oracle's internal physical address of a row (data file, block, and row-within-block), so accessing a row by ROWID means Oracle goes directly to the exact physical location — no further lookup is needed, making it the fastest possible access path. Primary key access and access via a unique index both first traverse the index's B-tree structure to find a matching ROWID before doing the actual table access by ROWID — they're a step slower because of that index-scan step. A full table scan is the slowest, since it reads every block in the table (or up to the high-water mark) checking each row against the filter, with no direct addressing at all.