reading a file in reverse order (bootom-top)

S

sahukar praveen

Hello,

I have a question.

I try to print a ascii file in reverse order( bottom-top). Here is the logic.

1. Go to the botton of the file fseek(). move one character back to avoid the EOF.
2. From here read a character, print it, move the file pointer (FILE*) to 2 steps back (using fseek(fp, -2, SEEK_CUR)) to read the previous character.

This seems to be ok if the file has a single line (i.e. no new line character). The above logic fails if it encounters a new line character and gets into an infinite loop priting a series of new-line character.

To fix this I checked for the character read and if it is new-line character, I move the file pointer by 3 steps (fseek(fp, -3, SEEK_CUR)) and now the logic works fine.

Can anyone please explain me why a this special consideration for a new-line character.

Many Thanks in advance.

Thanks
Praveen
 
J

John Bode

sahukar praveen said:
Hello,

I have a question.

I try to print a ascii file in reverse order( bottom-top). Here is the
logic.

1. Go to the botton of the file fseek(). move one character back to
avoid the EOF.
2. From here read a character, print it, move the file pointer (FILE*)
to 2 steps back (using fseek(fp, -2, SEEK CUR)) to read the previous
character.

This seems to be ok if the file has a single line (i.e. no new line
character). The above logic fails if it encounters a new line character
and gets into an infinite loop priting a series of new-line character.

To fix this I checked for the character read and if it is new-line
character, I move the file pointer by 3 steps (fseek(fp, -3, SEEK CUR))
and now the logic works fine.

Can anyone please explain me why a this special consideration for a
new-line character.

Many Thanks in advance.

Thanks
Praveen
--

Some platforms represent a newline with a carriage return and line
feed pair, rather than a single newline character, which is why you
need to back up three characters instead of two.

Instead of reading a single character at a time, you might want to
grab a chunk of characters into a buffer, then print from that buffer.
For one thing, it reduces the number of calls to fseek() (which may
be a fairly expensive operation), and you can just skip over the
characters you don't want to print.
 
M

Mark McIntyre

Hello,

I have a question.

I try to print a ascii file in reverse order( bottom-top).

A quicker way if you have enugh memory would be to malloc a block of
memory you think is large enough for the entire file, fread() the file
into it, go to the last byte and walk backwards through the block.
PLatform specific extensions would help you find how much memory you
needed.
 
D

Debashish Chakravarty

Mark McIntyre said:
A quicker way if you have enugh memory would be to malloc a block of
memory you think is large enough for the entire file, fread() the file
into it, go to the last byte and walk backwards through the block.
PLatform specific extensions would help you find how much memory you
needed.

The solution I could think of printing the file in reverse order
without attempting to know the size of the file was a recursive one.
But I think this solution might choke on large files. This recursive
solution is equivalent to pushing characters on a stack and then
popping them, I think reading more than one character at a time would
led to faster code.

#include <stdio.h>
#include <stdlib.h>

/*print the characters of the file in reverse order*/
void rev_file_c(FILE *fp)
{
int c;
c=fgetc(fp);
if(c==EOF) return;
rev_file_c(fp);
putchar(c);
}

#define MAX_LINE_SIZE 4096
/*print the lines of the file in reverse order*/
void rev_file_lin(FILE *fp)
{
char buf[MAX_LINE_SIZE];
char *s;
s=fgets(buf,sizeof buf,fp);
if(NULL==s)
{
if(ferror(fp))
{
perror(NULL);
exit(EXIT_FAILURE);
}
else return;
}
rev_file_lin(fp);
fputs(buf,stdout);
}

int
main(void)
{
FILE *fp;
fp=fopen("c:\\temp\\rev_file.c","r");
if(NULL == fp)
{
perror(NULL);
return 1;
}
rev_file_c(fp);
fclose(fp);
return 0;
}
 
C

CBFalconer

Debashish said:
The solution I could think of printing the file in reverse order
without attempting to know the size of the file was a recursive one.
But I think this solution might choke on large files. This recursive
solution is equivalent to pushing characters on a stack and then
popping them, I think reading more than one character at a time would
led to faster code.

Try this one. I would write a few things differently today, and
there is an insect in revstring, which won't get triggered here.

/* Routines to reverse a file, char by char. */
/* by C.B. Falconer, 19 Dec. 2001 */
/* Released to public domain. Attribution appreciated */

/* Known bugs - A file without an initial empty line */
/* will have one added. */

#include <stdio.h>
#include <string.h>
#include <stdlib.h>

#define MAXLN 256

typedef struct line {
char *ln;
struct line *next;
} line, *lineptr;

/* ======================================= */
/* reverse string in place. Return length */
size_t revstring(char *string)
{
char *last, temp;
size_t lgh;

lgh = strlen(string);
last = string + lgh; /* points to '\0' */
while (last-- > string) {
temp = *string; *string++ = *last; *last = temp;
}
return lgh;
} /* revstring */

/* ========================= */
/* Reverse file, end to end */
int main(void)
{
char buffer[MAXLN];
lineptr p, last;
size_t lgh;

p = last = NULL;
while (fgets(buffer, MAXLN, stdin)) {
lgh = revstring(buffer);
if (p = malloc(sizeof (line))) {
p->next = last;
if (p->ln = malloc(lgh + 1)) {
strcpy(p->ln, buffer);
last = p;
}
else {
free(p);
break;
}
}
else break;
}
p = NULL;
while (last) {
free(p);
fputs(last->ln, stdout);
p = last;
last = last->next;
}
if (p && ('\n' != p->ln[strlen(p->ln) - 1]))
fputc('\n', stdout);
free(p);
return 0;
} /* main */
 
M

Mantorok Redgormor

CBFalconer said:
Debashish said:
The solution I could think of printing the file in reverse order
without attempting to know the size of the file was a recursive one.
But I think this solution might choke on large files. This recursive
solution is equivalent to pushing characters on a stack and then
popping them, I think reading more than one character at a time would
led to faster code.

Try this one. I would write a few things differently today, and
there is an insect in revstring, which won't get triggered here.

/* Routines to reverse a file, char by char. */
/* by C.B. Falconer, 19 Dec. 2001 */
/* Released to public domain. Attribution appreciated */

/* Known bugs - A file without an initial empty line */
/* will have one added. */

#include <stdio.h>
#include <string.h>
#include <stdlib.h>

#define MAXLN 256

typedef struct line {
char *ln;
struct line *next;
} line, *lineptr;

/* ======================================= */
/* reverse string in place. Return length */
size_t revstring(char *string)
{
char *last, temp;
size_t lgh;

lgh = strlen(string);
last = string + lgh; /* points to '\0' */
while (last-- > string) {
temp = *string; *string++ = *last; *last = temp;
}
return lgh;
} /* revstring */

/* ========================= */
/* Reverse file, end to end */
int main(void)
{
char buffer[MAXLN];
lineptr p, last;
size_t lgh;

p = last = NULL;
while (fgets(buffer, MAXLN, stdin)) {
lgh = revstring(buffer);
if (p = malloc(sizeof (line))) {
p->next = last;
if (p->ln = malloc(lgh + 1)) {
strcpy(p->ln, buffer);
last = p;
}
else {
free(p);
break;
}
}
else break;
}
p = NULL;
while (last) {
free(p);
fputs(last->ln, stdout);
p = last;
last = last->next;
}
if (p && ('\n' != p->ln[strlen(p->ln) - 1]))
fputc('\n', stdout);
free(p);
return 0;
} /* main */


#include <stdio.h>
#include <stdlib.h>
#include <string.h>

static unsigned int byte_count(FILE *);
static void reverse_file(char *, unsigned int);

int main(int argc, char **argv)
{
char *ptr;
unsigned int size;
FILE *fp;

if(argc < 2)
printf("%s <file>\n", argv[0]);


if(!(fp = fopen(argv[1], "r")))
return EXIT_FAILURE;

size = byte_count(fp);

if(!(ptr = malloc(size+1)))
return EXIT_FAILURE;

ptr[size] = '\0';

fread(ptr, size, sizeof(char), fp);

fclose(fp);

reverse_file(ptr, size);

return EXIT_SUCCESS;
}


static unsigned int byte_count(FILE *fp)
{
int c;
unsigned int count = 0;

while((c = fgetc(fp)) != EOF)
++count;

rewind(fp);

return count;
}

static void reverse_file(char *base, unsigned int size)
{
char *end = &base[size-1];

if(*end == '\n') {
*end = '\0';
--end;
}

for( ; end >= base; --end)
if(*end == '\n') {
*end = '\0';
printf("%s\n", end+1);
}

printf("%s\n", base);

free(base);
}

I think my version might outperform yours.
fgets is kind of slow on large files.
 
C

CBFalconer

Mantorok said:
.... snip ...

I think my version might outperform yours.
fgets is kind of slow on large files.

I believe it doesn't do the same thing. In addition I doubt that
reading a file twice can be faster than reading it once.
 
V

Villy Kruse

I believe it doesn't do the same thing. In addition I doubt that
reading a file twice can be faster than reading it once.

Not to mention that reading the entire contents of a huge file into memory
is not exactly the best thing to do. But of course having gigabytes of
real memory available will surely help.


Villy
 
C

CBFalconer

Villy said:
Not to mention that reading the entire contents of a huge file
into memory is not exactly the best thing to do. But of course
having gigabytes of real memory available will surely help.

Inasmuch as my version also reads everything into memory, I didn't
consider that a valid objection :) However the response to
memory allocation failure is also different.
 
J

Joe Wright

CBFalconer said:
Inasmuch as my version also reads everything into memory, I didn't
consider that a valid objection :) However the response to
memory allocation failure is also different.
Sorry I'm late. Here's mine. It reads the file twice but only one line
at a time is ever in memory.

#include <stdio.h>
#include <stdlib.h>

#define MAX(a,b) ((a) > (b) ? (a) : (b))

int main(int argc, char *argv[]) {
FILE *in = stdin, *out = stdout;
int ch, i, lines = 0, len = 0, max = 0;
long *linpa;
char *buff;
if (argc > 1)
if ((in = fopen(argv[1], "r")) == NULL)
fprintf(stderr, "Can't open %s\n", argv[1]),
exit(EXIT_FAILURE);

if (argc > 2)
if ((out = fopen(argv[2], "w")) == NULL)
fprintf(stderr, "Can't make %s\n", argv[2]),
exit(EXIT_FAILURE);

while ((ch = fgetc(in)) != EOF) {
++len;
if (ch == '\n') {
++lines;
max = MAX(len, max);
len = 0;
}
}

fprintf(stderr, "There are %d lines in %s\n", lines, argv[1]);
fprintf(stderr, "The longest of which is %d characters\n", max);

rewind(in);
buff = malloc(max + 1);
linpa = malloc((lines + 1) * sizeof *linpa);
if (buff == NULL || linpa == NULL)
fprintf(stderr, "Can't allocate memory\n"), exit(EXIT_FAILURE);
i = 0;
linpa = 0;
while ((ch = fgetc(in)) != EOF)
if (ch == '\n')
linpa[++i] = ftell(in);

for (i = lines - 1; i >= 0; --i) {
fseek(in, linpa, SEEK_SET);
fgets(buff, max+1, in);
fputs(buff, out);
}
return EXIT_SUCCESS;
}
 
C

CBFalconer

Joe said:
Sorry I'm late. Here's mine. It reads the file twice but only one line
at a time is ever in memory.

It still doesn't do the same thing :) (besides reading the file
twice). Note that my version reversed the file, not just the line
order. I.e. every byte was reversed in order.
 
J

Joe Wright

CBFalconer said:
It still doesn't do the same thing :) (besides reading the file
twice). Note that my version reversed the file, not just the line
order. I.e. every byte was reversed in order.
Boo. That's not so hard at all.

.....

for (i = lines - 1; i >= 0; --i) {
fseek(in, linpa, SEEK_SET);
fgets(buff, max+1, in);
rev(buff);
fputs(buff, out);
}
return EXIT_SUCCESS;
}

void rev(char *src) {
char *des = src;
int tmp;
while (*des) ++des;
while (--des > src) {
tmp = *des;
*des = *src;
*src++ = tmp;
}
}

And mine is shorter than yours. (That really sounds wierd. Sorry.) :=)
 
J

Jeff Rodriguez

sahukar said:
Hello,

I have a question.

I try to print a ascii file in reverse order( bottom-top). Here is the
logic.

1. Go to the botton of the file fseek(). move one character back to
avoid the EOF.
2. From here read a character, print it, move the file pointer (FILE*)
to 2 steps back (using fseek(fp, -2, SEEK_CUR)) to read the previous
character.

This seems to be ok if the file has a single line (i.e. no new line
character). The above logic fails if it encounters a new line character
and gets into an infinite loop priting a series of new-line character.

To fix this I checked for the character read and if it is new-line
character, I move the file pointer by 3 steps (fseek(fp, -3, SEEK_CUR))
and now the logic works fine.

Can anyone please explain me why a this special consideration for a
new-line character.

Many Thanks in advance.

Thanks
Praveen
Here's a code example ripped directry from my "C Primer Plus" book:
Listing 13.5 The reverse.c Program
--snip--
/* reverse.c -- displays a file in reverse order */
#include <stdio.h>
#include <stdlib.h>
#define CNTL_Z '\032' /* eof marker in DOS text files */
#define SLEN 50
int main(void)
{
char file[SLEN];
char ch;
FILE *fp;
long count, last;

puts("Enter the name of the file to be processed:");
gets(file);
if ((fp = fopen(file,"rb")) == NULL)
{ /* read-only and binary modes */
printf("reverse can't open %s\n", file);
exit(1);
}
fseek(fp, 0L, SEEK_END); /* go to end of file */
last = ftell(fp);
for (count = 1L; count <= last; count++)
{
fseek(fp, -count, SEEK_END); /* go backward */
ch = getc(fp);
/* for DOS, works with UNIX */
if (ch != CNTL_Z && ch != '\r')
putchar(ch);
/* for Macintosh */
/* if (ch == '\r')
putchar('\n');
else
putchar(ch); */
}
putchar('\n');
fclose(fp);
return 0;
}
--snip--
This operates on the idea that your lines are no longer than 50 chars... but
that's easy to work around. This should give you the basic idea.

Jeff
 
C

CBFalconer

Jeff said:
.... snip ...
Here's a code example ripped directry from my "C Primer Plus" book:
Listing 13.5 The reverse.c Program
--snip--
/* reverse.c -- displays a file in reverse order */
#include <stdio.h>
#include <stdlib.h>
#define CNTL_Z '\032' /* eof marker in DOS text files */
#define SLEN 50
int main(void)
{
char file[SLEN];
char ch;
FILE *fp;
long count, last;

puts("Enter the name of the file to be processed:");
gets(file);
^^^^
Enough said. Don't buy that book. Burn it if you have it. Also
missing any fflush.
 
R

Richard Heathfield

Jeff said:
Here's a code example ripped directry from my "C Primer Plus" book:
Listing 13.5 The reverse.c Program

puts("Enter the name of the file to be processed:");
gets(file);

This is enough to condemn the book.
if ((fp = fopen(file,"rb")) == NULL)
{ /* read-only and binary modes */
printf("reverse can't open %s\n", file);
exit(1);
}
fseek(fp, 0L, SEEK_END); /* go to end of file */

If the file is binary, SEEK_END is not guaranteed to work. The Standard
says: "A binary stream need not meaningfully support fseek calls with a
whence value of SEEK_END."

<snip>
 
J

Jeff Rodriguez

Richard said:
Jeff Rodriguez wrote:







This is enough to condemn the book.




If the file is binary, SEEK_END is not guaranteed to work. The Standard
says: "A binary stream need not meaningfully support fseek calls with a
whence value of SEEK_END."

<snip>
"This should give you the basic idea. "

This snippet came from a chapter in the book before memory management, like I
said.. the basic idea. Dump all over the book if you wish, just keep in mind
that snippet is very out context and the reason I posted it here is convey the
basic idea of how to do what the OP wanted.

Jeff
 
R

Richard Heathfield

Jeff said:
"This should give you the basic idea. "

But it doesn't - or at least, it doesn't from the point of view of this
newsgroup.
This snippet came from a chapter in the book before memory management,
like I said.. the basic idea. Dump all over the book if you wish, just
keep in mind that snippet is very out context and the reason I posted it
here is convey the basic idea of how to do what the OP wanted.

It conveys a non-portable technique which might be appropriate in a
platform-specific newsgroup such as comp.os.msdos.programmer, but not in a
newsgroup such as this one, where we don't allow ourselves the luxury of
code that depends on non-portable tricks.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

Forum statistics

Threads
473,769
Messages
2,569,580
Members
45,055
Latest member
SlimSparkKetoACVReview

Latest Threads

Top