Composite Data Type
-single data type that is referenced from other data type
- Data type can be primitive or user-defined
- Record : a collection related items that have different data types
- Array : indexed collection of items with same data type
- List : indexed collection of items that can have different data types
- Class : gives the properties and methods for an object
Non-composite data type
- Single data type that does not involve a reference to another data type
- Usually built in to a programming language.
- Integer : stores a whole number
- Boolean : stores true or false
- Real : stores decimal number
- String : stores one or more character
- Pointer : whole number used to reference a memory location
Enumerated data type
- a user-defined non-composite data type with an ordered list of possible values
Pointer data type
- a user-defined non-composite data type used to reference a memory location
Why need user-defined data type ?
- no suitable data type is provided in language used
- Programmer need to specify new data type
- Referenced from one or more existing data type
- Extend the built-in data types
- That meet requirements of application
Changes of Mantissa/Exponent
More bits for mantissa - better precision
More bits for exponent - larger range
Why some binary representation lead to rounding errors ?
- some number cannot be represented by exact binary number
- More number of bits than available is needed to store the number
Problems occur due to change in the representation of floating-point number
- number calculated will change
- many numbers having same representation
- software may clash if not updated
Why floating point number is normalised?
- first 2 bits is opposite
Why binary numbers are stored in normalized form ?
- store maximum range of number in minimum number of bytes
- Minimise num of leading zero represented
- Maximising precision of the number for the given number of bits
- Enable very large/small number to be stored with accuracy
- Avoids possibility of many numbers having same representation
Problem occur when floating point number are not normalised
- Precision lost
- Redundant leading zeros in mantissa
- Bits loss off right hand end
- Multiple representation of a single number
Overflow
- occur in exponent of a floating-point number
- Exponent become too large to be represented using the number of bits available.
Underflow
- Following an arithmetic/logical operation
- number too small
- Cannot be represented by the number of bits available
Sequential
sequential
Random
direct
Serial
direct/sequential
Why serial file organisation ?
- high hit rate
- records are stored in chronological order
- Easy to append new record to EOF
- Allows reading of records in the order that they were taken
- No key fields need to be added
- No need sort data every time new data is added (order not important)
- Only a small file so searching only require a little processing
Why sequential file organisation ?
- high hit rate
- has key field to be used
- Data can be sorted
- Suitable for batch processing
Why serial file organisation use sequential access ?
– earliest record is accessed first
- And each successive records is read
- Until final reading has been accessed
Compare serial and sequential
- Both file records are stored one after the other
- And need to be accessed one after the other
- Serial file is stored in chronological order
- Sequential files are stored in the order of key field
- And stored in ordered records
- Serial files, new record is appended to the file
- Sequential files, new records are inserted in correct position.
How random file organisation work
- home locations are calculated
- Using a hashing algorithm on a key field
- If home location is free, insert record
- If a record cannot be stored at that location
- Then subsequent locations are searched
- Or use overflow method to find free location to store record
- If no free location available then record cannot be stored
How random file organisation’s record is found
- key field is hashed to get home location
- Compared ID to the home location
- If nothing stored, output msg “record not found”
- Record ID equal, record is found
- Record ID not equal, search overflow area
- Until record found or whole area searched
- If no record found error msg
Why random file organisation is appropriate
- Low hit rate
- speed of access
- No need search through all records
- Suitable for access to single record - Just used as a look up file
- No need for any serial or sequential processing
Describe the user-defined data type record
- A user-defined record is a composite data type
- It uses other data types in its definition to form a single new data type
- Data types are referenced may be primitive from a programming language/they may be other user-defined data type
- Include related items
- Include fixed number of items.
