Printable View
Hello,
I am writing a c program in to copy all the lines from a file to another file, only in this case, while copying from the original file,i need to skip a few lines.
What i mean is that a few lines from the oringinal file should not be copied to the new file.
I am using the fgets function.
What is the easiest solution to this?
When you read in a line you don't want, don't write it out.
fgets only takes the first line, you dont know how many lines are there in the text so i think you should write your own fgets function to do better operations.
Not:before calling this function; you should open your file, and after you called the function you should close the file.
you can do here, if there is an \n and after if there is unwanted words etc. , you may continue your loop.Code://***********MY FGETS FUNCTION**************
char *Tomgets(FILE *filex,char *filename)
{
char *p;
char c;
int i=0;
len=-1; //abc\0 == 0 1 2 3 (their position in array a[0] a[1]... )
for(i=0;c!=EOF;i++) // from starting of text file to END OF FILE ,the file length will calculated.Why? because if you want to save the content of the text file to an array, you will use this length to open the size of array
{
c=(fgetc(filex)); //gets all the letters in a loop
len++;
}
fclose(filex);
//-------------closed and reopened file to start fgetc from beginning to the end-------------
filex=fopen(filename,"r+"); //file opened for reading
p=(char *)malloc(len*sizeof(char));
for(i=0;i<len;i++) p[i]=(fgetc(filex)); //saves the content of the file to 'p' array.
//p[i]='\0'; you can do it to do better.
return p; //function returns to the text file content.
}
By what criteria... Content? Length? Marker? ???Quote:
Originally Posted by alienator
Just read the file in a loop...Quote:
I am using the fgets function.
What is the easiest solution to this?
It's not hard...Code:#define MAX_LINE_LENGTH 80 // change as necessary
char buffer[MAX_LINE_LENGTH]
// open input file
// make sure it's open
// open output file
// make sure it's open
while (fgets(buffer,MAX_LINE_LENGTH,infile))
{
if (! ExcludeLine(buffer)) // write exclusion routine like this
fputs(buffer,outfile);
}
// close input file
// close output file
1) fgets() reads the file one line at a time... got 25000 lines... use a loop.Quote:
Originally Posted by rac1
2) Nobody reads files character by character... that's just slow and silly
3) If the file is already open why on earth would you close it and re-open it,
4) If you did #3, why wouldn't you check if it actually opened?
5) if you want to know how big a file is...
6) In c you don't cast the return value of malloc()... if your compiler complains it's in C++ mode.Code:fseek(file,0,SEEK_END);
size = ftell(file);
fseek(file,0,SEEK_SET);
7) Your routine is horridly inefficient since it has to read the file twice... character by character, no less.
8) if you simply want to read the whole file into memory...
Code:char *buffer = malloc(size);
fread(buffer,1,size,file);
You missed a few:
9) c is uninitialised and could therefore easily happen to be initially equal to EOF causing the first loop to be skipped.
10) the resulting string that is returned is not null-terminated thus the caller has no way to determine the length of the string. Note that the malloc does not even allocate enough room for a null-terminator.
11) i is unnecessarily given the value of zero twice initially.
12) There is no check to see that the malloc succeeded.
13) After calling this function, the file handle that was passed in is invalid. Since filex is not passed out, the file is now stuck open and cannot be closed.
14) Somewhat poor indentation and use of whitespace.
I sure hope somebody learnt a few things here!
@ Common Tater
>>>When you read in a line you don't want, don't write it out.
That is right, but the thing is that i am dealing with duplicate lines here. I need to keep only 1 copy of the original but i need a way to distinguish in between.
Usually, getting rid of duplicate strings is easiest by sorting the strings, which will group the duplicates all together. Then in your input of the strings, you check the string behind it ONLY, to see if it's a duplicate. If it is, just don't write it out, and take in the next string - keep comparing until you reach a string that is not the same as your "reference" string.
Sometimes, you need to do this, but you also need to keep the strings in their original order - that's no problem. You can "sort" them, without actually moving them, by using an integer array (or pointer array), and swapping the integer index instead of swapping the actual strings. You wind up with the original data, unchanged, but now you can remove the duplicates, using the above technique, by referring to the strings THRU the index array (or pointer array).
It's harder to describe than it is to code it up, once you know how the logic works.
Interested?
If you have some spare memory (trade-off), a potentially more efficient way is to hash each string, and go through them one by one, and for each one, if the hash is already in a hash table, skip the line, otherwise add the hash to the hash table. Only keys are used in the hash table.
If the hash table is big enough and the hashing algorithm is good, this is O(n).
This is similar in concept to bucket sort.
I was thinking along similar lines to cyberfish... hashing is one technique, so is checksumming (similar but easier)... just make a checksum value for each string as you load it... stuff it into an array or linked list and check if there's any matches as you load more strings... The advantage is that it reduces 64+ byte strings down to 4 bytes... Checksums are very easily produced, just add up the character values in each string then multiply by the number of characters... Not perfect but unique enough to give you a 90% solution.
A checksum is really nothing but a poor hash. FNV-1a is about as simple yet far better and very widely used and documented.
Personally I tend to use a CRC for hashing.