我想要计算文件中的字符(在各种字符集中)并且我使用函数'mbtowc'来检测字符.我无法弄清楚为什么字符和结果值不同.这是我的例子:
char buf[BUFFER_SIZE + MB_LEN_MAX];
int fd = open ("chinese_test", O_RDONLY);
unsigned int bytes, chars;
int bytes_read;
bytes = chars = 0;
while((bytes_read = read(fd, buf, BUFFER_SIZE)) > 0) {
wchar_t wc_buf[BUFFER_SIZE], *wcp;
char *p;
int n = 0;
bytes += bytes_read;
p = buf;
wcp = wc_buf;
while((n = mbtowc(wcp, p, MB_LEN_MAX)) > 0) {
p += n;
wcp++;
chars++;
}
}
printf("chars: %d\tbytes: %d\n", chars, bytes);
Run Code Online (Sandbox Code Playgroud)
我用一些带有GB2312字符的文本测试函数,但字符和字节的值太大了.
我的测试返回 - >字符:4638 | bytes:17473但'wc'linux命令返回:chars:16770 | 字节:17473
为何如此区别?我做错了什么?
现在我已经有了这段代码,但结果仍然存在差异.
char buf[BUFFER_SIZE * MB_LEN_MAX];
int fd = open ("test_chinese", O_RDONLY), filled = 0;
unsigned int bytes, chars;
int bytes_read;
bytes = chars = 0;
while((bytes_read = read(fd, buf, BUFFER_SIZE)) > 0) {
wchar_t wc_buf[BUFFER_SIZE], *wcp;
char *p;
int n = 0;
bytes += bytes_read;
p = buf;
wcp = wc_buf;
while(bytes_read > 0) {
n = mbtowc(NULL, p, MB_LEN_MAX);
if (n <= 0) {
p++;
bytes_read--;
continue;
}
p += n;
bytes_read -= n;
chars++;
}
}
printf("\n\nchars: %d\tbytes: %d\n", chars, bytes);
Run Code Online (Sandbox Code Playgroud)
问题是你BUFFER_SIZE的文件大小chinese_test和字节对齐的组合wchar_t.作为证据,尝试大幅增加BUFFER_SIZE- 你应该开始得到你想要的答案.
发生的事情是你的程序适用于它收到的第一个文本块.但是,如果字符在第一个和第二个块之间分开,请考虑代码中发生的情况,如下所示:
| First Block | Second Block |
| [wchar_t] [wchar_t] ... [wchar_t] [wchar_t] ... |
| [1,2,3,4] [1,2,3,4] ... [1,2,3,4] [1,2,3,4] ... |
Run Code Online (Sandbox Code Playgroud)
您的代码将在第一个字符的第3个字节开始第二个块,并且不会被识别为有效字符.由于在找不到有效字符时mbtowc将返回-1,因此您的循环将立即结束并将为整个块计算零个字符.这同样适用于以下块.
编辑:
我注意到的另一个问题是你需要设置语言环境才能mbtowc正常工作.考虑到所有这些问题,我编写了以下内容,它返回与wc我相同的字符数:
#include <stdlib.h>
#include <stdio.h>
#include <locale.h>
int BUFFER_SIZE = 1024;
const char *DEFAULT_F_IN = "chinese_test";
struct counts {
int bytes;
int chars;
};
int count_block(struct counts *c, char *buf, int buf_size)
{
int offset = 0;
while (offset < buf_size) {
int n = mbtowc(NULL, buf + offset, MB_CUR_MAX);
if (n <= 0) {
break;
}
offset += n;
c->bytes += n;
c->chars++;
}
return buf_size - offset;
}
void get_counts(struct counts *c, FILE *fd)
{
char buf[BUFFER_SIZE];
c->bytes = 0;
c->chars = 0;
int bytes_read;
while((bytes_read = fread(buf, sizeof(*buf), BUFFER_SIZE, fd)) > 0) {
int remaining = count_block(c, buf, bytes_read);
if (remaining == 0) {
continue;
} else if (remaining < MB_CUR_MAX) {
fseek(fd, -remaining, SEEK_CUR);
} else {
perror("Error");
exit(1);
}
}
}
int main(int argc, char *argv[]) {
FILE *fd;
if (argc > 1) {
fd = fopen(argv[1], "rb");
} else {
fd = fopen(DEFAULT_F_IN, "rb");
}
setlocale(LC_ALL, "");
struct counts c;
get_counts(&c, fd);
printf("chars: %d\tbytes: %d\n", c.chars, c.bytes);
return 0;
}
Run Code Online (Sandbox Code Playgroud)