从CSV文件中读取整数值，如何仅获取记录的最后两个值？

Question

The sample record will be like this,

14/11/2014,Sh2345,423,10
12/12/2014,AV2345,242,20

From the above record I need only

423,10
242,20

The below code will give me all the row and Column count.

rowIndex = 0;
columnIndex = 0;
while(fgets(part,1024,fp) != NULL){
    token = NULL;

    while((token = strtok((token == NULL)?part:NULL,",")) != NULL){
        if(rowIndex == 0){ 
            columnIndex++;
        }
        for(idx = 0;idx<strlen(token);idx++){
            if(token[idx] == '\n'){ 
                rowIndex++;
                break;
            }
        }
    } 
}

Answer 1

If you want to use strtok which I believe is the right way to do this kind of things, since fscanf will be very problematic in case of invalid input, then I think this is the way:

rowIndex = 0;
while (fgets(part, sizeof part, fp) != NULL)
{
    char *token;
    size_t partLength;
    char *saveptr; // for strtok_r to store it's current state

    partLength = strlen(part);
    /* check if this is a complete line */
    if (part[partLength - 1] == '\n')
        rowIndex++;

    columnIndex = 0;
    token       = strtok_r(part, ",", &saveptr);
    while ((token = strtok_r(NULL, ",", &saveptr)) != NULL)
    {
        char *endptr;

        /* if columnIndex >= 1 then we are in the right columns */
        if (columnIndex >= 1)
            values[columnIndex - 1] = strtol(token, &endptr, 10);
        /* in case the conversion rejected some characters */
        if ((*endptr != '\0') && (*endptr != '\n'))
            values[columnIndex - 1] = -1; /* some invalid value (if it's possible) */
        columnIndex++;
    }
    /* if we have columnIndex == 3, then we've read the two values */
    if (columnIndex == 3)
        printf("(%d, %d)\n", values[0], values[1]);
    /* the last column will not be counted in the while loop */
    columnIndex++;
}

In case of very long lines, for which sizeof part is small enough to leave some , in between, you are going to need some different approach, but as long as the lines fit part you are ok.

To read the values into an array, maybe this could work:

int **fileToMatrix(const char *const filename, int *readRowCount, int *readColumnCount, int skipColumns)
{
    char  part[256];
    FILE *file;
    int   rowIndex;
    int   columnIndex;
    int   index;
    int **values;

    file = fopen(filename, "r");
    if (file == NULL)
        return NULL;
    values   = NULL; /* calling realloc, it behaves like malloc if ptr argument is NULL */
    rowIndex = 0;
    while (fgets(part, sizeof part, file) != NULL)
    {
        char *token;
        int **pointer;
        char *saveptr; // for strtok_r to store it's current state

        /* check if this is a complete line */

        pointer = realloc(values, (1 + rowIndex) * sizeof(int *));
        if (pointer == NULL)
            goto abort;

        values           = pointer;
        values[rowIndex] = NULL;
        columnIndex      = 0;
        token            = strtok_r(part, ",", &saveptr);

        while ((token = strtok_r(NULL, ",", &saveptr)) != NULL)
        {
            columnIndex += 1;
            /* if columnIndex > skipColumns - 1 then we are in the right columns */
            if (columnIndex > (skipColumns - 1))
            {
                int   value;
                char *endptr;
                int  *currentRow;
                int   columnCount;

                endptr = NULL;
                value  = strtol(token, &endptr, 10);
                /* in case the conversion rejected some characters */
                if ((endptr != NULL) && (*endptr != '\0') && (*endptr != '\n'))
                    value = -1;
                /*           ^ some invalid value (if it's possible) */
                columnCount = columnIndex - skipColumns + 1;
                currentRow  = realloc(values[rowIndex],  columnCount * sizeof(int));
                if (currentRow == NULL)
                    goto abort;
                currentRow[columnIndex - skipColumns] = value;
                values[rowIndex]                      = currentRow;
            }
        }
        /* the last column will not be counted in the while loop */
        columnIndex++;
        rowIndex++;
    }
    fprintf(stderr, "%d rows and %d columns parsed\n", rowIndex, columnIndex - skipColumns);
    fclose(file);

    *readRowCount    = rowIndex;
    *readColumnCount = columnIndex - skipColumns;

    return values;

abort:
    *readRowCount    = -1;
    *readColumnCount = -1;

    for (index = rowIndex - 1 ; index >= 0 ; index--)
        free(values[index]);
    free(values);

    fclose(file);
    return NULL;
}

void freeMatrix(int **matrix, int rows, int columns)
{
    int row;
    for (row = 0 ; row < rows ; row++)
        free(matrix[row]);
    free(matrix);
}

void printMatrix(int **matrix, int rows, int columns)
{
    int row;
    int column;
    for (row = 0 ; row < rows ; row++)
    {
        int *currentRow;

        currentRow = matrix[row];
        for (column = 0 ; column < columns ; column++)
            printf("%8d", currentRow[column]);
        printf("\n");
    }
}

int main()
{
    int **matrix;
    int   rows;
    int   columns;

    matrix = fileToMatrix("data.dat", &rows, &columns, 2);
    if (matrix != NULL)
    {
        printMatrix(matrix, rows, columns);
        freeMatrix(matrix, rows, columns);
    }

    return 0;
}

You should also note, that sometimes fields in a CSV file contain " or ' quotes, you might want to remove them from the tokens returned by strtok_r to avoid the failure of strtol .

Answer 2

int v1, v2;
while(fgets(part,1024,fp) != NULL){
    sscanf(part, "%*[^,],%*[^,],%d,%d", &v1, &v2);//skip 2 field
    //do stuff .. printf("%d,%d\n", v1, v2);
}

Answer 3

int CheckMatrix(int Matrix, int Checkrow, int Checkvalue) /* to check whether the area code existing in the matrix*/
 {
    int i,j;

   for(i=0;i<=Checkrow;i++)
       {
          if( Matrix[i][0]== Checkvalue)
             { 
                return i;
             }
          else
             {
               return -1;
              }
        }
 }

int **fileToMatrix(const char *const filename, int *readRowCount)
{
    char  part[256];
    FILE *file;
    int   rowIndex;
    int   index;
    int **values;
    size_t partLength;
    int v1,v2;
    int CheckValue;


    file = fopen(filename, "r");
    if (file == NULL)
        return NULL;
    values   = NULL; /* calling realloc, it behaves like malloc if ptr argument is NULL */
    rowIndex = 0;
    while (fgets(part, sizeof part, file) != NULL)
    {
        int **pointer;

        /* check if this is a complete line */

        pointer = realloc(values, (1 + rowIndex) * sizeof(int *));
        if (pointer == NULL)
            goto abort;
         partLength = strlen(part);
        /* check if this is a complete line */
        if (part[partLength - 1] == '\n')
        rowIndex++;
        sscanf(part, "%*[^,],%*[^,],%d,%d", &v1, &v2);//skip 2 field to get the Area code and Distance 

        CheckValue = CheckMatrix(Values,rowIndex,V1); //Call the function to check  whether the area code existing or not in the array

         If (CheckValue!=-1) // If existing the current distace will add to the existing and increase the count of areacode.
           { 
             Values[CheckValue][1]=Values[CheckValue][1]+V2;
             Values[CheckValue][2]=Values[CheckValue][2];

            }
        else // If not existing will add to the matrix as new entry.
            {
              Values[CheckValue][0]=V1;
              Values[CheckValue][1]=V2; 
              Values[CheckValue][2]=1;
            }   


    }

    return values;

abort:
    *readRowCount    = -1;
    *readColumnCount = -1;

    for (index = rowIndex - 1 ; index >= 0 ; index--)
        free(values[index]);
    free(values);

    fclose(file);
    return NULL;
}

void freeMatrix(int **matrix, int rows)
{
    int row;
    for (row = 0 ; row < rows ; row++)
        free(matrix[row]);
    free(matrix);
}

void printMatrix(int **matrix, int rows)
{
    int row;
    for (row = 0 ; row < rows ; row++)
    {

         for (column = 0 ; column < 4 ; column++)
            {   
               if (column==3)
               {
                  double Mean = double(matrix[row][1])/double(matrix[row][2]); /* To get the mean */
                  printf("%f",matrix[row][column];
               }  

                printf("%d |", matrix[row][column]);

            }

            printf("\n");
    }
}




# include <stdio.h>
int main()
{
    int **matrix;
    int   rows;

    matrix = fileToMatrix("data.dat", &rows);
    if (matrix != NULL)
    {
        printf("|AreaCode|Total Distace|Area Count|Mean");
        printf("------------------------------------------");
        printMatrix(matrix, rows);
        freeMatrix(matrix, rows);
    }

    return 0;
}