Opening gzipped files for reading in C without creating temporary files

I have some gzipped files that I want to read in C via fopen and fscanf. Is there anyway to do this without having to gunzip the files to temporary files?

Th开发者_JAVA技巧anks.

You can use libzlib to open the gzipped files directly.

It also offers a "gzopen" function that behaves similar to fopen but operates on gzipped files. However, fscanf would probably not work on such a handle, since it expects normal FILE pointers.

If popen is fair game, you can do it with fopen and fscanf:

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

int main(int argc, char *argv[])
{
  const char prefix[] = "zcat ";
  const char *arg;
  char *cmd;
  FILE *in;
  char buf[4096];

  if (argc != 2) {
    fprintf(stderr, "Usage: %s file\n", argv[0]);
    return 1;
  }

  arg = argv[1];
  cmd = malloc(sizeof(prefix) + strlen(arg) + 1);
  if (!cmd) {
    fprintf(stderr, "%s: malloc: %s\n", argv[0], strerror(errno));
    return 1;
  }

  sprintf(cmd, "%s%s", prefix, arg);

  in = popen(cmd, "r");
  if (!in) {
    fprintf(stderr, "%s: popen: %s\n", argv[0], strerror(errno));
    return 1;
  }

  while (fscanf(in, "%s", buf) == 1)
    printf("%s: got [%s]\n", argv[0], buf);

  if (ferror(in)) {
    fprintf(stderr, "%s: fread: %s\n", argv[0], strerror(errno));
    return 1;
  }
  else if (!feof(in)) {
    fprintf(stderr, "%s: %s: unconsumed input\n", argv[0], argv[1]);
    return 1;
  }

  return 0;
}

For example:

$ zcat file.gz
Every good boy does fine.
$ ./gzread file.gz
./gzread: got [Every]
./gzread: got [good]
./gzread: got [boy]
./gzread: got [does]
./gzread: got [fine.]

Do not use

sprintf(cmd, "zcat %s", argv[1]);
popen(cmd,"r");

to open .gz files. Properly escape argv[1] instead. You may otherwise end up with a vulnerability, especially when some injects an argument argv[1] such as

123;rm -rf /

It already helps to change the above instruction into

sprintf(cmd, "zcat \'%s\'",argv[1]);

You may also want to escape characters such as '\0', '\'', '\;' etc.

Newbie attempt at gzscanf():

#include <stdio.h>
#include <stdarg.h>
#include <zlib.h>

#define MAXLEN 256

int gzscanf(gzFile *stream, const char *fmt, ...) {
  /* read one line from stream (up to newline) and parse with sscanf */
  va_list args;
  va_start(args, fmt);
  int n;
  static char buf[MAXLEN]; 

  if (NULL == gzgets(stream, buf, MAXLEN)) {
    printf("gzscanf: Failed to read line from gz file.\n");
    exit(EXIT_FAILURE);
  }
  n = vsscanf(buf, fmt, args);
  va_end(args);
  return n;
}

You can use zlib and wrap it to a regular file pointer, this way you can use fscanf,fread,etc. transparently.

FILE *myfopen(const char *path, const char *mode)
{
#ifdef WITH_ZLIB
  gzFile *zfp;

  /* try gzopen */
  zfp = gzopen(path,mode);
  if (zfp == NULL)
    return fopen(path,mode);

  /* open file pointer */
  return funopen(zfp,
                 (int(*)(void*,char*,int))gzread,
                 (int(*)(void*,const char*,int))gzwrite,
                 (fpos_t(*)(void*,fpos_t,int))gzseek,
                 (int(*)(void*))gzclose);
#else
  return fopen(path,mode);
#endif
}

You can use zlib, but it will require you to replace your I/O calls to be zlib-specific.

you have to open a pipe to do this. The basic flow in pseudo code is:

create pipe // man pipe

fork // man fork

if (parent) {
    close the writing end of the pipe // man 2 close
    read from the pipe // man 2 read
} else if (child) {
    close the reading end of the pipe // man 2 close
    overwrite the file descriptor for stdout with the writing end of the pipe // man dup2 
    call exec() with gzip and the relevant parameters // man 3 exec
}

You can use the man pages in the comments for more details on how to do this.

It's quite simple to use zlib to open .gz files. There's a reasonable manual over at zlib.net.

Here's a quick example to get you started:

#include <stdio.h>
#include <zlib.h>

int main( int argc, char **argv )
{
    // we're reading 2 text lines, and a binary blob from the given file
    char line1[1024];
    char line2[1024];
    int  blob[64];

    if (argc > 1)
    {
        const char *filename = argv[1];
        gzFile gz_in = gzopen( filename, "rb" );  // same as fopen()

        if (gz_in != NULL)
        {
            if ( gzgets( gz_in, line1, sizeof(line1) ) != NULL )  // same as fgets()
            {
                if ( gzgets( gz_in, line2, sizeof(line2) ) != NULL )
                {
                    if ( gzfread( blob, sizeof(int), 64, gz_in ) == 64 )  // same as fread()
                    {
                        printf("Line1: %s", line1);
                        printf("Line2: %s", line2);
                        // ...etc
                    }
                }
            }
            gzclose(gz_in);  // same as fclose()
        }
        else
        {
            printf( "Failed to GZ-open [%s]\n", filename );
        }
    }
    return 0;
}

Remember to link with zlib, under UNIX gcc ... -lz