Suppose we are given a database with the following schema. Users User IDINTEGER,

Question

Suppose we are given a database with the following schema. Users User IDINTEGER, Name CHAR(30), Age INTEGER, ReviewCount INTEGER) Businesses (BusinessID INTEGER, BName CHAR(30), City CHAR(20), State CHAR(2) Checkins (BusinessIDINTEGER, Weekdays INTEGER, Weekends INTEGER) Reviews (ReviewID INTEGER, UserIDINTEGER, BusinessID INTEGER, Stars REAL) Reviews (UserID) is a foreign key referring to Users (UserID). Reviews (BusinessID) is a foreign key referring to Businesses (BusinessID). Checkins (BusinessID) is a foreign key referring to Businesses (BusinessID) A page is 8 kB in size. The RDBMS buffer pool has 10,000 pages, all of which are usable. Initially, the buffer pool is empty. The relation instances have the following statistics. Assume there are no NULL values. Each integer or real is 8B, and each character is 1B (so as an example CHAR(20) is 20B). Additionally, the record id of each tuple is 8B. Relation Number of Pages Number of Tuples 75,684 10m. Users 41,504 5mm Businesses 19,532 5mm Checkins Reviews 488,282 100m

Explanation / Answer

M P = G F (M 1) P

P =( G F – M)/( M 2 )

where F is a fudge factor used to reflect the overhead from both the hash table and other structures (more about the fudge factor will be discussed in Section 6.7). This formula indicates that the total input data is processed as one resident partition (occupying M P frames for in-memory aggregation), and P spilled partitions (each will fit into memory using M 1 frames). The above formula appeared in [21] for joins; we adapt it here for aggregation, so it uses the result set size G instead of the input size R,

Related Questions

Navigate

• Browse (All)
• Browse S
• Subjects

---

---

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.