IBM Mainframe Forum

[Greenhouse]

Hi,

Which way is most efficient to compare two files for line equality?

[enrico-sorichetti]

from a practical or an algorithmic point of view

ever heard of ISPF superc, or HLASM toolkit superc ???

[NicC]

SUPERC, COMPAREX, other utilities for file compare. Possibly even SORT.

[Greenhouse]

I meant using language constructs without external tools.

[enrico-sorichetti]

the question as posed does not make any sense
kind way of telling that the question is ...

what do You want to compare...
are the files sorted or not
what kind of changes do You want to identify
do You want to compare at line level or also at block level
... and so on and so on

googling for ( for example ) file compare algorithms or two files match
You will find lots of links which will help You to ask the question in a better way

and at last how does this relate to C/C++ languages only ???

[dick scherrer]

Hello,

I meant using language constructs without external tools.

Nearly anything can be written but most organizations are not willing to pay to have code written that already exists. . .

Is this for your learning?

You need to post a bit of sample data from 2 files and the results you want when your code executes using the sample input data.

[Greenhouse]

The whole story is that i wrote some REXX and then just for my learning I try to rewrite it in C++.

In REXX I have line level comparison of two (sorted) files and report all difference with related line number.

What I wanted to do in C++ is to at least preform line level comparison and at first difference to report it and compare next pair.

As I thought load module will be faster then interpreted program, but it wasn't.

Select all

string f_res = "OK ";                                   
                                                         
ifstream fprep;                                         
fprep.open(prep_fname.c_str(), ios_base::in);           
                                                         
if(!fprep.is_open())                                     
   throw E(prep_fname += " file cannot be open", 8);     
                                                         
ifstream fexec;                                         
fexec.open(exec_fname.c_str(), ios_base::in);           
                                                         
if(!fexec.is_open())                                     
   throw E(exec_fname += " file cannot be open", 8);     
                                                         
string pline;                                           
string eline;                                           
                                                         
while(!fprep.eof())                                     
{                                                       
   if(fexec.eof())                                       
   {                                                     
      f_res = "NOK";                                     
      cout << "WORK CMPR " << tstid                     
         << " Different work files sizes.\n";           
      break;                                             
   }                                                     
                                                         
   getline(fprep, pline);                               
   getline(fexec, eline);                               
                                                         
   if(pline != eline)                                   
   {                                                     
      f_res = "NOK";                                     
      cout << "WORK CMPR " << tstid                     
         << " Difference between lines of work files.\n";
      break;                                             
   }                                                     
}
fprep.close();
fexec.close();
             
return f_res;

[dick scherrer]

Hello,

Suggest you look at the amount of cpu resource used. The rexx probably takes more cpu time.

The reason both run in similar amounts of elapsed time is probably because both are limited by the speed of the dasd. . . You may need to run a higher volume of test data to see much of a difference.