Question

Write a program that reads in a continuous stream of strings representing a line of CSV data in the format "NAME,AGE,EMAIL,DOB". Implement a function check_csv which verifies that each input string matches this format by ensuring that: • There are 4 tokens in the string corresponding to the 4 properties above. • The second token MUST be a number. • The fourth token should be in the format MM-DD-YYYY (hint: tokenize this as well). The function should return 0 if the CSV string is valid, and 1 otherwise. Your program should read CSV strings until the string "END" is entered as input. Print all valid strings to stdout followed by a line indicating how many invalid strings were entered. • You can use a character array with a buffer size of 1024 for reading each line using fgets. • You may not use any library specifically for parsing CSV. • Save your code as prob5.c.

Answer 1

#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <ctype.h>
int isleap(int year) //return 1 if leap year
{
  return (((year % 4 == 0) && (year % 100 != 0)) || (year % 400 == 0));
}
int isvaliddate(int d,int m,int y) //return 1 if date is valid
{
  int month[12]={31,28,31,30,31,30,31,31,30,31,30,31};
  if(isleap(y))
    month[1]=29;
  if(y>=1900 && y<=9999)//check year
  {
    if(m>=1 && m<=12) //check month
    {
      if(d>=1 && d<= month[m-1])
        return 1;
      else
        return 0;
    }
    else
      return 0;
  }
  else 
    return 0;
}
int check_csv(char *data)
{
  int isvalid=0; // let string is valid
  char dummy[100]; // dummy string to copy data
  strcpy(dummy,data); // strcpy to copy a string to other
  char* token = strtok(dummy,","); // tokenizing string
  char *attr[4]; // to store 4 tokens
  int i=0; 
  while (token != NULL)
  { 
    attr[i++]=token; //assign token in array
    token = strtok(NULL,","); //further token the string
  }
  //checking if attr[1] is number or not
  int count=0,flag=0;
  for(int i=0;i<strlen(attr[1]);i++)
  {
    if(isdigit(attr[1][i]))
      count++;
    if(attr[1][i]=='.')//for a float
    {
      flag=1;
    }
  }
  if(flag==0 && count==strlen(attr[1])) //valid case
    ;
  else if(flag==1 && count==strlen(attr[1])-1)
    ;
  else
    isvalid=1;
  //check date is valid or not
  char dummy2[100]; //to copy date value
  strcpy(dummy2,attr[3]); // strcpy to copy a string to other
  char* date_token = strtok(dummy2,"-"); // tokenizing string
  char *date[3]; // to store 3 tokens
  i=0; 
  while (date_token != NULL)
  { 
    date[i++]=date_token; //assign token in array
    date_token = strtok(NULL,"-"); //further token the string
  }
  if(strlen(date[0])!=2)
    return 1;
  if(strlen(date[1])!=2)
    return 1;
  if(strlen(date[2])!=5)//since space is also counted in it 
    return 1;
  // converting string to int
  int dd,mm,yyyy;
  sscanf(date[0],"%d",&mm);
  sscanf(date[1],"%d",&dd);
  sscanf(date[2],"%d",&yyyy);
  //check if date is valid or not by calendar
  if(!isvaliddate(dd,mm,yyyy))
    return 1;
  return isvalid;
}

int main()
{
        char *fname="data.csv"; //change file name
  //printf("Enter file name: ");
  //scanf("%[^\n]%*c", fname);
  FILE *file;
        file = fopen(fname, "r");//open file in read mode
        int i = 0;
  char line[1024]; //store line
  int inv=0; //count invalid lines
        while (fgets(line, 1024, file))
  {
    if(strcmp(line,"END")==0)//if End appear break input stream
      break;
    else
    {
      if(!check_csv(line))// if valid print
        printf("%s",line);
      else
        inv++;
    }
  }
  printf("\nTotal no. of invalid lines are: %d\n",inv);
  return 0;
}

Input: CSV file

Output: