1

friends! Programming language is C++. I have a byte matrix

unsigned char Map[Height][Width];

I initialize it with zeros. Then I read byte matrix from text file. Of course, text file can be larger, than my matrix is. In this case I must read information from file in such way: enter image description here

Text file has extra information. I don't need it. In another case Matrix can be larger then information within text file:

enter image description here

In this case program must read all the information from file. The part of matrix that didn't get information from file is already initialized. It's OK.

What is the best way to read information from file to the matrix in my case? I tried to use fgets, but it reads all information from file in consecutive way - byte after byte. But I don't need to read extra bytes. Of course, I can read file byte after byte and check counter. But I am sure that this is'nt best solution. Does a better solution exist?

The formatting of my text file is'nt relevant here. I read information from file like bytes field.

Improve this question
8
  • 4
    Why aren't you using the C++ ifstream? You could easily wrap that in a loop or other construct that handles the exceptional cases. Commented Jan 17, 2012 at 18:58
  • 3
    The formatting of your text file is relevant here. You'd want to include that in your question. Commented Jan 17, 2012 at 18:58
  • 2
    Why can't you just chop off the extra bytes from each line if they are too long? Commented Jan 17, 2012 at 18:58
  • 2
    If your coding C++ (not C), I'd suggest using the classes from <fstream> for file IO, as opposed to the C-style functions in <cstdio>. They're a lot easier to use! Commented Jan 17, 2012 at 18:59
  • Sorry, I am not familiar with <fstream> library. Commented Jan 17, 2012 at 19:04

3 Answers 3

Reset to default
1

Assuming your text files look something like this:

3254352
6536543
8875687
4315254
5345435
1212122

the solution with standard library streams would be something like this:

#include <fstream>
#include <string>

int main()
{
    std::ifstream matrix_file("matrix.txt");

    const size_t Width = 5;
    const size_t Height = 5;
    unsigned char Map[Width][Height];

    size_t line_count = 0;
    std::string line;

    while (std::getline(matrix_file, line) && line_count < Height) {
        line.resize(Width);
        for (size_t i = 0; i < line.length(); ++i)
            Map[i][line_count] = line[i];
        ++line_count;
    }
    if (line_count < Height) {
         // the file didn't have enough lines to fill our matrix,
         // so we'll need to fill the rest with zeroes here
    }
}

The line.resize(Width); line above will automatically put null characters in the string when Width is greater than the length of the line in the file. Also, don't forget to check if stream was opened succesfully before you try to read from it (fstream::is_open(), for example). I skipped that check for brevity.

Improve this answer
Sign up to request clarification or add additional context in comments.

6 Comments

is it possible to read number of characters = "Width" without resizing?
Does resulting "line" variable contain 13,10 bytes (end of line)?
It is, but then you're back to the less elegant, char by char, way of reading. And no, the new-line characters are discarded by getline().
So, there is no elegant way to read matrix without reading of extra bytes?
It can be done, but why are you concerned about this? Is memory a problem?
|
1

If you can read all the lines with fgets(), you can copy the i-th line character by character to the i-th row of your table. Stop when you encounter the end of the line or when you have copied Width characters. Be careful that fgets() might leave a NL character at the end of the line that you might want to remove (NL + CR on Windows).

Repeat the above procedure until you have read Height lines or reached the end of the file.

My solution is basically C (because you say you are not familiar with the C++ functions). You should use standard C++ libraries (fstream, string) if you want a real C++ implementation.

#include <cstdio>
#include <cstring>

#define Height (...)
#define Width  (...)

// I assume that each input line in the file can contain at most width * 3 characters.
// Three extra characters for NL, CR, 0.
// Change this if you expect the file to contain longer lines.
#define BUFFER_WIDTH (Width * 3 + 3)

unsigned char Map[Height][Width];

unsigned char line[BUFFER_WIDTH];

// Remove CR, NL at the end of the line.
void clean_line(char *line)
{
    int len = strlen(line);

    while (len > 0 && (line[len - 1] == '\n' || line[len - 1] == '\r'))
    {
        line[len - 1] = '\0';
        len--;
    }
}

void read_table(const char *filename)
{
    FILE *fp = fopen(filename, "r");

    int row = 0;
    while (!feof(fp) && row < Height)
    {
        fgets(line, BUFFER_WIDTH, fp);
        clean_line(line);

        int len = strlen(line);
        int rowLen = len > Width ? Width : len;
        for (int col = 0; col < rowLen; col++)
        {
            Map[row][col] = line[col];
        }

        row++;
    }

    fclose(fp);
}
Improve this answer

4 Comments

Yes, you are quite right. But I'd like to know better solution of this problem.
Thank you for help. But It's better to read file byte after byte in such case. It is more simple solution. I have found some examples of ifstream usage. It is quite clear, as for me. I'll try to use it.
In C you can use fgetc() instead of fgets().
Yes, I know it. "Better" means - to develop program in such way (it's just example): fgets(line,WIDTH,file) //I get a line without "end of line" symbols. Then I read next line without reading of extra symbols in the same way fgets(line,WIDTH,file) - and so on. May be I should perform fseek(next_line) to go to a next line. But I don't want to read extra characters.
1

Read the entire file into a buffer:

FILE *fp;
long len;
char *buf;
fp=fopen("thefileyouwanttoread.txt","rb");
fseek(fp,0,SEEK_END); //go to end
len=ftell(fp); //get position at end (length)
fseek(fp,0,SEEK_SET); //go to beg.
buf=(char *)malloc(len); //malloc buffer
fread(buf,len,1,fp); //read into buffer
fclose(fp);

Copy the file into your byte array, check to see which is bigger to determine how to copy:

char *startByteArr;
unsigned char Map[Height][Width];

startByteArr = &Map[0][0];
if (len > Height*Width){
    memcpy(startByteArr,buf,Height*Width);
else {
    memcpy(startByteArr,buf,len);
}

This assumes that the first dimension is the same though. To account for varying width in the file, you could change the memcpy like:

char *startByteArr;
char *EndLineFile;
int lineLengthFile;

EndLineFile = strstr (buf,"\r\n");
lineLenghtFile = (int)(EndLineFile-buf);

unsigned char Map[Height][Width];

startByteArr = &Map[0][0];

int i;

if (lineLengthFile > Width){
    for(i = 0; i < Height;i++){
        memcpy(startByteArr+i*Width,buf+i*lineLengthFile,Width);
    }
else {
    for(i = 0; i < Height;i++){
    memcpy(startByteArr+i*Width,buf+i*lineLengthFile,lineLengthFile);
    }
}

I believe that would be the fastest way, just grab the whole file into memory in one read, then memcpy the segments you need to a byte array.

Improve this answer

2 Comments

It needs extra memory. I think that it isn't the best solution of the problem.
The above needs a buffer to store the file in memory, but when optimizing, one often chooses between speed and memory size, my solution tries for speed, as memory is plentiful in most present day computers. I doubt you loading anything > 500 MB into your Map.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.