Parsing a csv file having comma separated value in single column using C

try this

if (file == NULL){ perror("Not able to open the file\n"); exit(EXIT_FAILURE); } while (fgets(line,1000, file)!=NULL){ if( 3==sscanf(line, "%d,\"%499[^\"]\",%d", &arg1, arg2, &arg3) || //pattern 1 3==sscanf(line, "%d,%499[^,],%d", &arg1, arg2, &arg3)){ //pattern 2 printf("Value to insert Key:%d,Value:%s,Height:%d\n",arg1,arg2,arg3); inserth(inode,arg1,arg2,arg3); } else { fprintf(stderr, "invalid format\n"); exit(EXIT_FAILURE); } } 

One solution is to implement your own strtok that respects double-quotes. strtok by itself is a pretty simple to implement:

char * tokenise( char *str, const char *delim ) { static char *next = NULL; if( str ) next = str; if( !*next ) return NULL; str = next; while( !strchr( delim, *next ) ) next++; if( *next ) *next++ = 0; return str; } 

Now, that's the general case. You only care about commas, and in the case of double quotes, you only care about double quotes:

char * tokenise( char *str ) { static char *next = NULL; if( str ) next = str; if( !*next ) return NULL; str = next; if( *str == '"' ) { str++; next++; while( *next && *next != '"' ) next++; if( *next == '"' ) *next++ = 0; } while( *next && *next != ',' ) next++; if( *next ) *next++ = 0; return str; } 

This is pretty naive, but should do the trick. It detects " in the first character, removes it and then scans to the next ". It doesn't handle escaped quotes, whitespace between CSV fields, or syntax errors (such as non-comma characters appearing after a closing quote -- it will drop them) but you get the idea.

First of all, for a csv file, if a comma comes inside double quotes, it ususally means the comma should be ignored. This means that every row has only 3 values (3 columns).

Since you have only two types of rows, one with double quotes and one without double quotes, as simple if-else statement would help here

I am using strsep instead of strtok. Also I am reading from the file line by line.

#include <string.h> #include <stdio.h> int main() { FILE *file; char *splitline; char * line = NULL; size_t len = 0; ssize_t read; file=fopen(filename,"r"); if (file == NULL) { printf("Not able to open the file\n"); } while ((read = getline(&line, &len, file)) != -1) { if(strstr(line,"\"") != NULL) { printf("%s\n",line); splitline=strsep(&line,","); printf("%s : %s\n",splitline,line); line = line+1; // avoiding first doublequotes("), you may use strsep twice instead splitline = strsep(&line,"\""); printf("%s : %s\n",splitline,line); line = line+1; // removing comma and rest is the 3rd entry printf("%s",line); } else { //Routine code that expects two commas and three values } } }