Arrays

[pbhuter]

I have several text files of columns of data. Some have three columns, some have four. All of the columns have headers on them. I need to be able to break each test file into an array based on the columns. I need to be able to access the data (doubles) in the columns. I will be using a regular expression to figure out which column has the data I need, based on the header in the array.

Thanks in advance.

[hamsterman]

fstream file("...");

string temp;
getline(file, temp);//read the first line with headers

stringstream ss(headers);
int count = 0, index = -1;

for(; ss >> temp; count++)
  if(/*temp is the string you wanted*/) index = count;
//now count shows how many columns you have
//and index shows which one you want
//index = -1 indicates failure

double val;
for(int i = 0; i < index; i++) file >> val;//skip the first several columns

vector<double> values;//you can use an array if you know how long the lists are.
//You'll want to make the following loop into an for loop in that case.

while(file){//while input hasn't failed (due to end of file or a string instead of a number)
   file >> val;
   values.push_back(val);//read and push a value you need into the vector.
   //if you want an array, the two lines above would turn into
   //file >> values[counter]; where counter is the variable of the for loop I mentioned.

   for(int i = 0; i < count-1; i++) file >> val;//skip the remaining entries of this line and some of the other.
}

Something like that. I'm assuming the file doesn't have empty entries. That would make it more complicated..

[pbhuter]

hamsterman - Thank you. Here is what I have (based on your code above):

		while (!datFile.eof())
		{
			string temp;
			getline (datFile, temp);

			stringstream ss(headers);

			int count = 0, index = -1;

			for (; ss >> temp; count++)
			{
				if (temp...)
				{
					index = count;
				}
			}

			double val;
			for (int i = 0; i < index; i++)
			{
				datFile >> val
			}

			vector<double> values;

			while (datFile)
			{
				datFile >> val;
				values.push_back(val);
			}

stringstream ss(headers) doesn't work - what is "headers"?

I want the first column and the column with the data I want, so how do I modify lines 19-22 for this purpose?

I obviously need something more on line 12 where I have temp... - I need to do a search (using regular expressions?) to see if that header contains "eu" or "EU" - Do you know how to do this, or should I post again in the forum?

I must apologize for this post - I'm used to having the Qt infrastructure to help me out, but I don't have it on this computer.

Thanks, again!

[Gaminic]

Could you maybe post up a few lines of example data file? Would make it easier to explicitly point to parts of it.

[pbhuter]

Sure!:

%b02_a08 b02_a08_counts b02_a08_eu
-0.998350 16 2.252804
-0.998292 16 2.252804
-0.998233 16 2.252804
-0.992525 16 2.252804
-0.992467 16 2.252804
-0.992408 16 2.252804
-0.986825 16 2.252804
-0.986767 16 2.252804

This is the one with three columns. I want the first column and the last one (the one that contains "eu".

%b10_a43 b10_a43_gps b10_a43_counts b10_a43_eu
-59.990000 14340000.000000 0 0.000000
-59.979700 14340100.000000 0 0.000000
-59.969500 14340100.000000 0 0.000000
-59.959300 14340100.000000 0 0.000000
-59.949000 14340100.000000 0 0.000000
-59.938800 14340100.000000 0 0.000000
-59.928500 14340100.000000 0 0.000000
-59.918300 14340100.000000 0 0.000000
-59.908100 14340100.000000 0 0.000000
-59.897800 14340100.000000 0 0.000000
-59.887600 14340100.000000 0 0.000000
-59.877300 14340200.000000 0 0.000000
-59.867100 14340200.000000 0 0.000000

This one has four columns, again, I want the first column and the stuff with "eu".

[pbhuter]

Also, since I am loading multiple data files, I will need multiple vectors (not arrays, as this topic states...). Here is my code so far:

	string name;
	
	fstream inFile;
	fstream datFile;
	inFile.open ("input.txt");

	if (!inFile)
	{
		cerr << "File open failure\n";
		exit(EXIT_FAILURE);
	}

	while (!inFile.eof())
	{
		getline (inFile, name);

		datFile.open(name);

		if (!datFile)
		{
			cerr << "File open failure\n";
			exit(EXIT_FAILURE);
		}

		while (!datFile.eof())
		{
			string temp;
			getline (datFile, temp);

			stringstream ss(headers);

			int count = 0, index = -1;

			for (; ss >> temp; count++)
			{
				if (temp...)
				{
					index = count;
				}
			}

			double val;
			for (int i = 0; i < index; i++)
			{
				datFile >> val;
			}

			vector<double> values;

			while (datFile)
			{
				datFile >> val;
				values.push_back(val);
			}

	}

See above notes for questions about "temp..." and "datFile >> val".

I want to have multiple vectors (one for each file). The above code will loop over all the "datFile" names listed in "inFile.txt", and create a vector for each one.

Thanks!

[hamsterman]

what is "headers"?

Sorry. That was supposed to be temp.

You have way too many if (!inFile) and similar checks in your code. It might be a good idea to check once that the file was opened successfully and you really need the one I had in my code, but there is no point in having all of the rest.
Edit: didn't notice you had two files.. Line 25 is still not needed though.

About if(/*temp is the string you wanted*/), I assumed you had the regex part figured out. Well, you could do as you said and get yourself a regex library like Boost.Regex. Though I think with C++11 there is a standard one..
Or you could simply use http://cplusplus.com/reference/string/string/find/ which would be sufficient for your current simple needs.

About the whole thing, if you don't need just one line, I suggest you read all of the data and then pick out what you want. That way it will be less of a mess. You'll need a vector of strings for headers and a vector of vectors of doubles for the columns. Ask if you can't figure out how that would work.

[pbhuter]

hamsterman:

I'm going to be going through the data, doing some interpolation, and eventually generating +/- 3-sigmas for all the data points. I'll play with C++ vectors and see what I come up with. If I can't solve it, I'll post back.

Thanks.