Computing – Programming
File Organisation
Serial File:
 In a serial file, the data is stored in the order in which it arrives.
 New data is simply added to the end of the file. This is called appending the data.
 If an item of data is to be deleted, this will create a gap in the file. This can be removed by recreating the file: copying all the records in the file, apart from the record to be deleted, into a new file and using this file to replace the original file.
 To search for data in a serial file, you will need to start from the first item and search for each item in turn until you find the record you are looking for. If the item you are searching for is not in the file, you will need to search every item in the file before you can be certain that the item is not found.
Sequential File
 In a sequential file, the data are stored according to a key field in the data. The key field is a field in each record that can be used to identify the record. For example, if each student were given a StudentlD, then a sequential file of students will be stored in order of the Student ID.
 The addition of an item to a sequential file requires the new item to be inserted within the existing records. This is usually achieved by recreating the file. All the records up to the correct point of insertion are copied to a new file. The new record is then appended in the correct position and the remaining records are copied into the new file. The new file then replaces the old file.
 If an item is deleted this would create a gap in the sequence. This can be prevented by recreating the file in the same manner as deleting an item from a serial file.
 Searching for an item in a sequential file is usually similar to searching for an item in a serial file. You start from the first item and you search each item in turn until you reach the item you are searching for. However, if you reach an item that has a higher value for the key than the item you are searching for, then you do not need to carry on searching till the end of the file to know that the item is not found.
Indexed sequential File
 An indexed sequential file is a sequential file where the data is arranged according to a key field in the record, but which also has an index that allows the records to be found directly.
 This is useful in files where sometimes it is necessary to process all records in sequential order, and sometimes it is necessary to access individual records. For example, if the program is printing a register of the students in a student file then it will access the data sequentially. However, if you want to look up the details of an individual student, then the index is used to find the required record quickly.

Direct-access files:
 If the records in a sequential file have a fixed length then it may be possible to read the contents of any record in the file directly if you know its position in the file.
 For example if the size of a record is 40 bytes, then the records in the file will be at positions 0, 40, 80, 120, etc.
 To find the tenth record, it is not necessary to go through all the records in the file. The position or 'address' of the record can be calculated using the following formula:
Address of Record number n = beginning of file + ((n-1) * Size of record)
 In this case the tenth record will be at position 360.
Random files:
 In all the file types we have discussed so far, the records are physically stored next to each other either serially or sequentially.
 A random file allows the data to be stored anywhere in a dedicated section of a disk. A calculation called a hash algorithm is performed on a field in the record to be stored.
 The result of this calculation then becomes the address. As a result, the records can appear randomly scattered across the disk. However, finding the data is extremely quick as long as you know the key field and the hash algorithm. There is no need to refer to the other data in the file.
 As a result, random files are very useful for large databases where individual records are often looked up individually.
Estimating the size of a file:
 It is possible to estimate the size of a data file using the number of records in the file and the sizes of the fields in each record.
 The procedure is as follows:
 Determine the size, in bytes, of each field in the record (for a string field, use the largest size allowed).
 Add the field sizes to calculate the size of one record.
 Multiply the size of one record by the number of records in the file.
 Add 10% to the result, for additional storage that will be needed to manage the file on the disk (we call this overheads).
 This will give you an estimate of the size of the file in bytes. If necessary, it can be converted into kilobytes (kB) by dividing by 1024, and then converted again into megabytes (MB) by dividing again by 1024.
 However, for the purposes of estimating, dividing by 1000 instead of 1024 would give a reasonably adequate result.
 For example, suppose a school had 1800 students and kept their details in a file of records structured as follows:
RECORD Student Name : String Gender : Character Age : Integer
END RECORD  The total size of each record is 23 bytes.
 The size of 1800 records is therefore 23 x 1800 = 41,400 bytes.
 Adding 10% for overheads gives us 41,400 x 1.1 = 45,540 bytes.
 Dividing this by 1000 gives us an estimate of 45.5 kB.
Preparing the file:
 Most languages will require the program to make the operating system aware that a file is about to be used.
 This is often done using an OPEN command. When the file is opened it is often necessary to specify how the file is going to be used (e.g. read only, read and write data, write data only, append to existing data, direct access, etc.). This is called the access mode.
 It protects the data in the file from being corrupted, for example, by conflicting instructions using the same file.
 For this reason, it is good practice to open a file for as little time as necessary. When the operation is over, the file should be closed. This releases the file so that it can be used by another part of the program, or another program.
Reading data from a file:
 Languages provide different methods for reading data. The simplest way is to read the data from the file one line at a time. This is useful for searching through a serial file.
 In this case it is necessary to test, before reading each line, that the end of the file has not been reached. For random and direct access files, you need to be able to read the data that is at a given address in the file.
Writing data to a file:
 You also need to be able to write the data into a file that has been correctly opened.
 This can include replacing (updating) the data in a direct-access file, or adding data to the end of a serial file (appending).
 Inserting data into a sequential file will require all subsequent records to be moved to make space for the data to be inserted. An alternative method is to create a new file and copy all the records from the present file into it, inserting the new data in the file at the correct location.
 The new file then replaces the old file. This is called merging the data into the sequential file.
File management:
 Programming languages also provide facilities for copying, moving, deleting files, and finding out if a file exists.
 The following algorithm uses several of the file operations discussed to delete a record from a sequential file:
BEGIN
Open SourceFile for reading Create NewFile Open NewFile for writing REPEAT Read a record from SourceFile IF it is not the record you want to delete THEN Write the record to NewFile END IF UNTIL end of sourcefile Close SourceFile Close NewFile Delete SourceFile Rename NewFile to the name of the SourceFile
END