Random access file using fstream

[Muthuveerappan]

Hi,

I wanted to do the following using fstream:
1) Write 3 employee records into a new file - was able to do this
2) Then the first character at the beginning of the 2nd record needs to be replaced with 'Z' - Wasn't able to this, I wasn't able to position seekp() correctly.

Pls let me know where I went wrong and why.

Given below is the program:

#include <iostream>
#include <string>
#include <fstream>
using std :: string;
using std :: ostream;
using std :: cout;
using std :: endl;
using std :: fstream;
using std :: ios;

class EmpRecord
{
    public:
        EmpRecord(const int pEid, const string pName, const double pSalary);
        static int getSize();
   
        const int eid;
        const string name;
        const double salary;
};

ostream& operator << (ostream& pOutStream, const EmpRecord& pEmpRecord);

int main()
{
    EmpRecord emp1(1, "abcd", 100.20),
              emp2(2, "efgh", 231.31),
              emp3(3, "ijkl", 200.93);

    fstream destFile("outfile.txt", ios :: out);
    
    system("clear");
    
    cout << emp1 << emp2 << emp3;
    cout << "size of EmpRecord = " << EmpRecord :: getSize() << endl;
    
    destFile << emp1 << emp2 << emp3;

    destFile.seekp(EmpRecord :: getSize() - 1);
    destFile.put('Z');
    
    destFile.close();
    
    return(0);
}

EmpRecord :: EmpRecord(const int pEid, const string pName, const double pSalary)
    : eid(pEid), name(pName), salary(pSalary)
{}

int EmpRecord :: getSize()
{
    int sizeInBytes = sizeof(int) + sizeof(string) + sizeof(double) + 2 * sizeof('\t') + sizeof('\n');
    return(sizeInBytes);

}

ostream& operator << (ostream& pOutStream, const EmpRecord& pEmpRecord)
{
    pOutStream << pEmpRecord.eid << '\t'
               << pEmpRecord.name << '\t'
               << pEmpRecord.salary << '\n';

    return(pOutStream);
}

Thanks,
Muthu

[Philip Nicoletti]

1) in getSize() you are using sizeof(string) ... that is a constant regardless of how many
characters are in the string.

2) You can use seekp() for text files to get to specific records ... but the records need
to be the same size.

3) On non-Posix systems (read Windows), I would open the file in binary mode
to prevent carriage return / linefeed translations.

[Lindley]

It's impossible to predict bytes in a file if every record doesn't have precisely the same size. Since you're using formatted output and a variable-length string, your records definitely don't all have the same size here.

[Muthuveerappan]

Thanks a lot Philip Nicoletti and Lindley, I will try that as per your suggestions and get back on this.

Sorry for the delayed response.