Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
25 Cards in this Set
- Front
- Back
3 Diff Types of Indexes |
-B+ tree, Hash, & Bitmap |
|
B+ tree |
levels of ranges, getting to final range allows you to access the actual data |
|
Hash |
compresses data into smaller number to store, uses that number to access data (can cause collisions -- 2 or more elements with the same hash number) |
|
Bitmap |
works with and and or’s, usually takes the element and splits and assigns numbers based on attributes |
|
External Constraints |
foreign key -- updating is a challenge (might update one without the other) |
|
Internal Constraints |
not null (person’s name that attends the church), primary key (person’s SSN) |
|
Why Views? |
-data independence → you can rename columns and it doesn’t matter -simplification -security -performance |
|
ACID |
Atomic, Consistent, Isolated, Durable |
|
Atomic |
all of it happens or none of it happens |
|
Consistent |
if its consistent before, then it’s after; it may deter in the middle, but it will finish off |
|
Isolated |
what the husband does not interfere with what the wife does |
|
Durable |
if the system crashes at any time, once a transaction is committed, a transaction is permanent |
|
Problems with Concurrency |
-manipulating common data -lost update -uncommitted dependency -inconsistent retrieval |
|
Lock Types |
-read/shared -read/write or exclusive |
|
Transaction |
a collection of operations that must be processed as one unit of work |
|
Lock Reliability |
the smaller the locks, the smaller the overhead and greater concurrency |
|
Recovery Transparency |
DBMS restores database to consistent state after failure |
|
Concurrency Transparency |
Users perceive database as single user system |
|
Objective of Concurrency |
maximize throughout while preventing interference |
|
3 Interference Problems |
-lost update -uncommitted dependency -inconsistent retrieval |
|
How to Avoid Deadlocks |
-detection -avoidance -time-out |
|
1st way to get serializability in DBMS= 2 Phase Locking Protocol |
-before accessing an item, one must acquire the appropriate lock -if one cannot acquire a lock, one waits -after releasing one lock, a transaction can’t acquire any new locks (or wait until end of transaction) -growing and shrinking phases |
|
2nd way to get serializability in DBMS= Optimistic Approaches |
-assumes conflicts are rare -doesn’t use locks -checks before commit or after read/write for conflicts - if problem, abort |
|
Types of Database Failures |
transaction initiated power lost hardware failure |
|
When should changes be recorded? |
-immediately - write it whenever (might have to undo stuff) -deferred - can’t write until commit (only have to redo) |