DBMS - File Structures

Question 1

File

Answer 1

named collection of records stored. Usually resident on secondary storage units ex. Disk

Question 2

File Processing

Answer 2

Management of these files ex. creation insertion, updating and deletion

Question 3

Two main modes of processing

Answer 3

• Batch (as a whole)
• Online (one record at a time)

Question 4

What to provide for in File Structures and Processing

Answer 4

• Design requirements
• Understanding a DBMS
• Comparison and comprehension of software
• Data Reorganization
• Query optimization purposes

Question 5

File Processing trains designers with

Answer 5

analytical and engineering techniques

Question 6

Need for file processing

Answer 6

• A Computation’s input and output can be larger than any main memory can hold
• A considerable amount of data sharing is required
• Persistent storage between bouts of file processing needed

Question 7

Mechanism for data access from secondary devices

Answer 7

Different from main memory

Question 8

Granularity of disk access

Answer 8

Cluster/Sector/Block based

Question 9

Data on storage devices can get

Answer 9

Lost/Damaged from physical or logical events

Question 10

Files placed on

Answer 10

Persistent storage devices. Secondary storage is not directly addressable for computation.

Question 11

IO instructions between main and secondary storage

Answer 11

Specialized

Question 12

Access time and magnitude

Answer 12

Has much larger magnitude than main memory accesses

Question 13

Background insights

Answer 13

• The cost of secondary storage is decreasing
• The disk performances are improving (good cost performance ratio)
• One can easily replace a disk unit with another

Question 14

Batch

Answer 14

comes in spans ex. end of day/week

Question 15

Online/Transaction Model

Answer 15

Data processed as soon as it arrives. Activity load spread and users are catered at the same time.

Question 16

Access Method

Answer 16

Program that implements selection criteria and navigation to manage data in a file or on behalf of an application program.

Question 17

Types of Files

Answer 17

Master
Reference
Transaction
Temporary
Archive
Log Files

Question 18

Master File

Answer 18

Used as an authority in a given job and is relatively permanent but can change.

Question 19

Reference File

Answer 19

Usually with an array of data value pairs. A reference file’s intention is a more static master file, with less data content and access to its records must be the fasdtest

Question 20

Transaction File

Answer 20

Used for a specific part of the business process. Processed with master and reference file(s).

Question 21

Temporary File

Answer 21

Acts as a dynamic, structured, or transient space. Data usually created and purged within the span of the same procedure

Question 22

Archive Files

Answer 22

Files record user and system actions, and historically relevant but now inert data

Question 23

Log Files

Answer 23

Support transaction processing requirements

Question 24

File protection

Answer 24

Loss of data due to programming or physical damage (but delimiting unlawful access)

Question 25

File security

Answer 25

Protection of data files from system operational hazards. Involves standard and severe crashes

Question 26

Standard crash

Answer 26

Coined for situations where data is lost while in main memory

Question 27

Severe crash

Answer 27

when secondary storage system's integrity/activity are affected.

Question 28

Records

Answer 28

Details that concern a single item present in a data file are grouped into a record. Data file can have a large number of records.

Question 29

Record formats

Answer 29

- Structured (Known set of attributes/fields) - Semi-structured (flexible set of attributes)

Question 30

Structured record size requirement

Answer 30

May be fixed or variable depending on fields.

Question 31

Fields/Attributes

Answer 31

- Each data detail in record is a field. Each field has a name, datatype and size - Some fields are special ex. key fields - Size depends on medium and file organization

Question 32

Organizing fields in record

Answer 32

Fixed length, Length+data pair, Field and record delimiters, Field name plus data pair

Question 33

File Organization (serial)

Answer 33

- Records have no value based ordering - Can have any size, most serial files have intrinsic temporal order - Usually is an input into some processing operation ex. sorting ex. Serial Search

Question 34

File Operations

Answer 34

Open file, read next char, read line, open file for output, write line

Question 35

File Organization (Sequential)

Answer 35

- Records ordered on one of its fields (key) ascending or descending - Can have any size - Uses sorted nature to drive another procedure Ex. Performance just the same as serial file

Question 36

File Organization (Direct)

Answer 36

- Given a record's key value, translate to an address of record in data file - Record access performance probably the best - Records placement in data file, independent of record's order by a key field - Accommodates free lists Deletion: Records reused on deletion (linked list) use the free list. Ex. Direct RRN access

Question 37

File Organization (Indexed Sequential)

Answer 37

- Data file records are maintained in key sequence but access direct record access is supported. - Hybrid between sequential and direct access files - Slower than sequential to read consecutive records - Slower than direct files for direct access of records - Best to handle both direct and sequential access of records Ex. Direct by key, Sequential (on key)

Question 38

Back of the envelope calculation

Answer 38

Estimate of size, effort etc. 1. Determine main parameters that describe and affect the problem 2. Derive a relationship that given known set-up values and parameters of the process required describes the scenario 3. Tone values for data volumes and most typical access methods and apply the relationship above to yield an estimated solution.