|
FILE SYSTEMS AND THE DATABASE CONCEPT- dasd = direct access storage devices. normally spinning disks
- overview of data access methods
- sequential files: data sorted into desired sequence and processed in sequence: difficult to access related records in more than two files: difficult to perform ad hoc inquiries
- random access files: (ISAM, VSAM): direct access to a specific record using pointers
- difficult to access multiple records related to a single record
- definition: pointer-: address of a record-: sector+offset
- history (sequential access-:direct access-:physical database-:logical database)
- early dp systems: automation of bookkeeping functions -: "data processing systems"
- program design-: mimic the existing manual process
- paper files -: computer files [ terminology: files, records, fields ]
- data files -: data for each application system.
- Example: customer file, salesRep file, product file, sales file
- sequential access-: mag tape, punched cards, paper tape, vsam disk files, batch processing
- sequential files: how the data are ordered is important: REWIND
- COBOL: sequential processing of files to produce -: updated files and printed reports
- Files -: Sorted Files -: PROCESS -: updated files/reports : sort/merge
- SALES -: sort by customer id-: sorted SALES -: update CUSTOMER file -: print statements
- SALES -: sort by REP-: sorted SALES-: update REP file-: print commission summary
- Batch runs made at periodic intervals (daily, monthly, yearly)
- problem: ad hoc query, OLTP, limited to two files, files must be sorted
- solution: direct access to each record using a key
- problem: multiple related records: example: list all sales to customer x. can't do it.
- problems with file oriented business systems:
- data redundancy-:time delayed and entry inconsistency, data definition inconsistency
- data control-: homonyms and synonyms, no common data structure or dictionary
- data manipulation-:inadequate (multiple record problem)
- expensive programming and maintenance-: heavy interdependence between file and application
- the database concept. data dictionary. database administrator.
- hierarchical database: supports access to multiple records related to a single record: restricted to hierarchical data relationships: dependent on predefined physical pointers.
- network database: supports hierarchical and non-hierarchical data relationships: dependent on predefined physical pointers.
- relational database: supports all logical relationships: independent of implementation technique
|