Objectives:

1
Chapter 5 Managing Files of Records

• Objectives
• To get familiar with
• Record Access
• File organization and file access
• Advanced file organizations

2
Outline

• Record access
• More about record structure
• File access and file organization
• Beyond record structures
• Portability and standardization

3
Record Access

• Record Keys
• Primary key
• unique, data-less, unchanging
• e.g Social Security Number
• Secondary key
• other fields to search generally not unique
• e.g., name, address

4
Record Access (contd)

• Sequential Search
• Read through the file record by record, looking
  for a record of a particular key.
• Count the number of READs -- assume all READs are
  equally costly
• The average number of READs needed for
  sequentially searching a file of n records is
  n/2. The sequential search has order O(n), i.e.,
  linear order.
• Placing several records together as a block and
  reading a block at a time.
• Sequential search with block is still of linear
  order.
• When to use sequential search
• Searching text file for patterns
• Small files -- files with few records
• Files that rarely need to be searched
• Searching on secondary keys and expecting a large
  number of matches

5
Record Access (contd)

• Direct access
• Directly position at the beginning of the record
  with a single SEEK( ) call and then read the
  record
• Order O(1) -- constant, independent of the number
  of records
• Need to know the position of the record before
  hand.
• Assume the relative record number (RRN) is known.
• The first record in a file has RRN 0 the second
  has RRN 1, and so on.
• If the records in a file is of fixed length, say
  r bytes, the byte offset of the record with RRN n
  is
• byte offset n ? r.
• Sometimes, an index is stored in a separate file.

6
More about Record Structure

• Fixed record length
• fixed field length
• variable field length
• A head record is often placed at the beginning of
  the file to hold some general information about a
  file.
• The header record has a different structure from
  the data records.
• The header record may include a count for the
  number of records in the file, the length of the
  data records, the date and time of the files
  most recent update, etc.
• A file can be self-describing by keeping file
  structure information in a header, e.g.,
• the name for each field,
• the width of each field, and
• the number of fields per record.

7
File Access and File Organization A Summary

• File organization depends on what use you want to
  make of the file.
• Since using a file implies accessing it, file
  access and file organization are intimately
  linked.
• Example though using fixed-length records makes
  direct access easier, if the documents have very
  variable lengths, fixed-length records is not a
  good solution the application determines our
  choice of both access and organization.

8
Beyond Record Structure

• Abstract data models for file access
• Headers and self-describing file
• Metadata data about data
• Color raster images and FITS files
• Store metadata with data
• Mixing object types in one file
• tagged file
• Representation-independent file access
• logical file vs. physical file
• Extensibility
• for new file formats

9
Portability and Standardization

• Factors affecting portability
• Differences among operating systems
• Differences among languages
• Differences in machine architectures
• Achieving portability
• Agree on a standard physical record format and
  stay with it
• Agree on a standard binary encoding for data
  elements
• Number and text conversion
• Standard intermediate format, such as XDR
• File structure conversion
• File system differences
• Unix and portability
• the dd utility