Thread: Code to search a file and take out a column

Originally Posted by iMalc

Do you even need to write a program for this!
If this is just for a one-off, you can just import it into Excel and then copy-n-paste the columns out into other files. (if you have access to Excel)
Or you could just attach the file to your post here and I'll split it for you.

Thanks guys for your help, im trying to put something together now, ive just never used some of the stuff used for doing this before.

And yes I do need a program lol, i have a very large number of files to go through and do this for.

Here is something that I have been using, just read everything in with the first function and use the second function to separate the data by column.

LINE is a global typedef of vector<string> used to store each line of a CSV file, so the very first line will hold the column titles, which is used by the second function to separate the columns.

Code:

// Reads the datafile into vector stats
vector<LINE> readFile(const char *filename)
{
    string line;
    int pos;

    ifstream fin(filename);

    if(!fin.is_open())
	{
		cout << "Failed to open file" << endl;
		system("EXIT");
	}

    while(getline(fin,line))
	{
		LINE ln;

        while( (pos = line.find(',')) >= 0)
		{
			string field(line.substr(0,pos));
			line = line.substr(pos+1);
			ln.push_back(field);
		}
		stats.push_back(ln);
	}
    return stats;
}

// Parses the vector into rows of information by column and returns the row
string GetColData(int row_number, string ColName)
{
    LINE &columnNames = stats[0];

    // find the column with the above column name
    for(int column = 0; column < columnNames.size(); column++)
    {
        if( ColName == columnNames[column] )
        {
            // get data for this column
            LINE &row = stats[row_number];
            return row[column];
        }
    }
    return "";
}

I almost forgot, stats is also global, but this was just a temporary fix, in the final prog it is not global:

Code:

vector<LINE> stats;

Originally Posted by m37h0d

normally i'd agree but for the purposes of the OP who cares?

ok ok ok - i care. it's a simple modification too.

also i just remembered that . is the concatenation operator, not +

Originally Posted by matsp

A lot of the "what's best/easiest" depends on EXACTLY what the format is. If the file is simply something like this:

Code:

1 2 3
4 5 6

then using the method described by Salem will be the simples (and perfectly adequate as long as you don't need to care about "idiot entered rubbish into the file" or "file got destroyed by attempt to copy it that failed"etc) method.

If you have a file that contains "extra decorations" (e.g. quotes or commas), then you will need to come up with a more advanced method of reading the data, as you need to somehow deal with those "decorations" - the method described by Salem won't work DIRECTLY, but if you know the format, and deal with it appropriately, it can be solved by a method SIMILAR to Salems - again with the provision that if you need to deal with error conditions, it gets a bit harder (handling errors ALWAYS adds more work than just doing the stuff that doesn't deal with errors - in fact, more of the code I write deal with "if(error)..." than the code that actualy does "real work").

--
Mats

Ok i see, so the numbers in my file are in a format like this 0.28403E+17 so ill need to correct for that.. right?

I didn't write the code that comes up with these numbers so im not sure of why they come out that way. I do know that it was written in FORTRAN if that makes a difference.

Originally Posted by coletrain

Hey guys, I am needing to make a code that can search through different files(that have 3 columns of numbers in them) and put all of the first columns in one file and all the 2nd in a different one.. etc.

Ive had some basic C++ classes but im not quite sure how to go about this. Is this even possible in C++? Looking for some pointers to get started.

assuming input files f1 f2 f3 and colums c1 c2 c3
output files o1 o2 o3

Read a line from f1 f2 f3
write
f1.c1 f2.c1 f3.c1 to o1
f1.c2 f2.c2 f3.c2 to o2
f1.c3 f2.c3 f3.c3 to o3
Repeat for each line of input files.