简单动态字符串
redis构建了SDS数据结构,主要对c字符串的一种补充.
sds.h
结构如下:
struct __attribute__ ((__packed__)) sdshdr8 {
//记录buf数组中已使用字节的数量
//等于SDS保存字符串长度,这里缓存了长度,将获取长度复杂度从O(N)降低到O(1)
uint8_t len; /* used */
//记录buf尚未分配字节数量
uint8_t alloc; /* excluding the header and null terminator */
unsigned char flags; /* 3 lsb of type, 5 unused bits */
//保存字符串,保留一字节存储空字符'\0'
char buf[];
};
与C字符串对比,SDS具备如下优点:
- 通过len记录字符串长度,从而将获取长度复杂度从O(N)降低到O(1)
- 通过free记录尚未分配的策略,减少内存重分配次数,从而增加性能
- 字符串以空字符串作为字符串结尾,所以不能存储二进制数据,而SDS则通过len判断字符串是否结束,从而达到了支持二进制目的.
- 通过遵循字符串以空字符结尾的规则,从而可以重用string.h很多函数.
通过free记录尚未分配的策略,主要有如下策略:
- 预分配
对SDS修改时,若free空间够,则写入.否则比较SDS修改之后的长度是否>1M,若<,则free=len,则具体占用空间为free+len+1(保存空字符串)
若>1M,则分配1M空间,同上. - 惰性空间
主要优化字符串缩减操作,当执行字符串缩减时,并不会释放内存,而是free记录释放个数.
源码
/**
*
* 初始化一个sds
* @param init 字符串
* @param initlen 字符串长度
* @return 返回sds的buf部分
*
* * 复杂度
* T = O(N)
*/
sds sdsnewlen(const void *init, size_t initlen) {
void *sh;
sds s;
//获取类型
char type = sdsReqType(initlen);
/* Empty strings are usually created in order to append. Use type 8
* since type 5 is not good at this. */
//若为SDS_TYPE_5,则默认设置类型为SDS_TYPE_8
if (type == SDS_TYPE_5 && initlen == 0) type = SDS_TYPE_8;
//获取类型长度
int hdrlen = sdsHdrSize(type);
unsigned char *fp; /* flags pointer. */
//zmalloc 不初始化所分配的内存
sh = s_malloc(hdrlen+initlen+1);
// 内存分配失败,返回
if (sh == NULL) return NULL;
if (init==SDS_NOINIT)
init = NULL;
else if (!init)
//将分配的内存全部初始化为 0
memset(sh, 0, hdrlen+initlen+1);
//将s指针偏移到buf
s = (char*)sh+hdrlen;
fp = ((unsigned char*)s)-1;
//初始化sh
switch(type) {
case SDS_TYPE_5: {
*fp = type | (initlen << SDS_TYPE_BITS);
break;
}
case SDS_TYPE_8: {
//将sh转成type对应的数据类型
SDS_HDR_VAR(8,s);
sh->len = initlen;
sh->alloc = initlen;
*fp = type;
break;
}
case SDS_TYPE_16: {
SDS_HDR_VAR(16,s);
sh->len = initlen;
sh->alloc = initlen;
*fp = type;
break;
}
case SDS_TYPE_32: {
SDS_HDR_VAR(32,s);
sh->len = initlen;
sh->alloc = initlen;
*fp = type;
break;
}
case SDS_TYPE_64: {
SDS_HDR_VAR(64,s);
sh->len = initlen;
sh->alloc = initlen;
*fp = type;
break;
}
}
//如果有指定初始化内容,将它们复制到buf中
if (initlen && init)
memcpy(s, init, initlen);
//buf最后一位填充 '\0'
s[initlen] = '\0';
//返回buf部分
return s;
}
/* Create an empty (zero length) sds string. Even in this case the string
* always has an implicit null term. */
/**
*
* @return 返回一个保存了空字符串的sds
* * 复杂度
* T = O(1)
*/
sds sdsempty(void) {
return sdsnewlen("",0);
}
/* Create a new sds string starting from a null terminated C string. */
/**
* 根据给定字符串,init sds
* @param init
* @return
*
* * 复杂度
* T = O(N)
*/
sds sdsnew(const char *init) {
size_t initlen = (init == NULL) ? 0 : strlen(init);
return sdsnewlen(init, initlen);
}
/* Duplicate an sds string. */
/**
* 复制一个sds
* * 复杂度
* T = O(N)
* @param s
* @return
*/
sds sdsdup(const sds s) {
return sdsnewlen(s, sdslen(s));
}
/* Free an sds string. No operation is performed if 's' is NULL. */
/**
* 释放sds
* * 复杂度
* T = O(N)
* @param s
* @return
*/
void sdsfree(sds s) {
if (s == NULL) return;
//释放
s_free(
//获取sdshdr起始偏移
(char*)s-
//根据数据类型获取sdshdr具体长度
sdsHdrSize(
//获取flag,也就是当前存储的数据类型
s[-1]));
}
/* Set the sds string length to the length as obtained with strlen(), so
* considering as content only up to the first null term character.
*
* This function is useful when the sds string is hacked manually in some
* way, like in the following example:
*
* s = sdsnew("foobar");
* s[2] = '\0';
* sdsupdatelen(s);
* printf("%d\n", sdslen(s));
*
* The output will be "2", but if we comment out the call to sdsupdatelen()
* the output will be "6" as the string was modified but the logical length
* remains 6 bytes. */
void sdsupdatelen(sds s) {
size_t reallen = strlen(s);
sdssetlen(s, reallen);
}
/* Modify an sds string in-place to make it empty (zero length).
* However all the existing buffer is not discarded but set as free space
* so that next append operations will not require allocations up to the
* number of bytes previously available. */
/**
* 不释放buf空间情况下, 重置sds保存的字符串
* * 复杂度
* T = O(1)
* @param s
*/
void sdsclear(sds s) {
sdssetlen(s, 0);
s[0] = '\0';
}
/* Enlarge the free space at the end of the sds string so that the caller
* is sure that after calling this function can overwrite up to addlen
* bytes after the end of the string, plus one more byte for nul term.
*
* Note: this does not change the *length* of the sds string as returned
* by sdslen(), but only the free buffer space we have. */
/**
* 对 sds 中 buf 的长度进行扩展,确保在函数执行之后,
* buf 至少会有 addlen + 1 长度的空余空间
* (额外的 1 字节是为 \0 准备的)
*
* * 复杂度
* T = O(N)
* @param s
* @param addlen
* @return
*/
sds sdsMakeRoomFor(sds s, size_t addlen) {
void *sh, *newsh;
//获取sds可用长度
size_t avail = sdsavail(s);
size_t len, newlen;
char type, oldtype = s[-1] & SDS_TYPE_MASK;
int hdrlen;
/* Return ASAP if there is enough space left. */
//可用长度>所需长度,则无需扩容
if (avail >= addlen) return s;
//获取sds占用空间长度
len = sdslen(s);
//获取之前类型的sh偏移
sh = (char*)s-sdsHdrSize(oldtype);
//s最少需要的长度
newlen = (len+addlen);
//根据新长度,为 s 分配新空间所需的大小
if (newlen < SDS_MAX_PREALLOC)
// 如果新长度小于 SDS_MAX_PREALLOC
// 那么为它分配两倍于所需长度的空间
newlen *= 2;
else
// 否则,分配长度为目前长度加上 SDS_MAX_PREALLOC
newlen += SDS_MAX_PREALLOC;
//获取新的数据类型
type = sdsReqType(newlen);
/* Don't use type 5: the user is appending to the string and type 5 is
* not able to remember empty space, so sdsMakeRoomFor() must be called
* at every appending operation. */
if (type == SDS_TYPE_5) type = SDS_TYPE_8;
hdrlen = sdsHdrSize(type);
if (oldtype==type) {
//重新分配长度
newsh = s_realloc(sh, hdrlen+newlen+1);
//分配失败返回NULL
if (newsh == NULL) return NULL;
//返回新的sds
s = (char*)newsh+hdrlen;
} else {
/* Since the header size changes, need to move the string forward,
* and can't use realloc */
//再次重新分配一个sdshdr
newsh = s_malloc(hdrlen+newlen+1);
//分配失败返回NULL
if (newsh == NULL) return NULL;
//将原来的sdr复制到重新分配的sdshdr上
memcpy((char*)newsh+hdrlen, s, len+1);
//释放原来的sh
s_free(sh);
//记录新的sdr偏移
s = (char*)newsh+hdrlen;
//设置新的数据类型
s[-1] = type;
//设置长度
sdssetlen(s, len);
}
//重新设置可分配长度
sdssetalloc(s, newlen);
return s;
}
/* Reallocate the sds string so that it has no free space at the end. The
* contained string remains not altered, but next concatenation operations
* will require a reallocation.
*
* After the call, the passed sds string is no longer valid and all the
* references must be substituted with the new pointer returned by the call. */
/**
* 缩减空间,分配的空间仅够存sds+1长度
* @param s
* @return
*/
sds sdsRemoveFreeSpace(sds s) {
void *sh, *newsh;
char type, oldtype = s[-1] & SDS_TYPE_MASK;
int hdrlen, oldhdrlen = sdsHdrSize(oldtype);
//获取sds占用空间长度
size_t len = sdslen(s);
//获取可用空间
size_t avail = sdsavail(s);
//获取sh偏移
sh = (char*)s-oldhdrlen;
/* Return ASAP if there is no space left. */
//无需分配
if (avail == 0) return s;
/* Check what would be the minimum SDS header that is just good enough to
* fit this string. */
//根据使用长度,获取对应类型
type = sdsReqType(len);
hdrlen = sdsHdrSize(type);
/* If the type is the same, or at least a large enough type is still
* required, we just realloc(), letting the allocator to do the copy
* only if really needed. Otherwise if the change is huge, we manually
* reallocate the string to use the different header type. */
//如果类型相同,或分配的类型过大
if (oldtype==type || type > SDS_TYPE_8) {
//重新分配
newsh = s_realloc(sh, oldhdrlen+len+1);
if (newsh == NULL) return NULL;
s = (char*)newsh+oldhdrlen;
} else {
//否则手动分配
newsh = s_malloc(hdrlen+len+1);
if (newsh == NULL) return NULL;
memcpy((char*)newsh+hdrlen, s, len+1);
s_free(sh);
s = (char*)newsh+hdrlen;
s[-1] = type;
sdssetlen(s, len);
}
//设置分配长度
sdssetalloc(s, len);
return s;
}
/* Return the total size of the allocation of the specified sds string,
* including:
* 1) The sds header before the pointer.
* 2) The string.
* 3) The free buffer at the end if any.
* 4) The implicit null term.
*/
/**
*
* @param s
* @return 分配字节数
* * 复杂度
* T = O(1)
*/
size_t sdsAllocSize(sds s) {
size_t alloc = sdsalloc(s);
return sdsHdrSize(s[-1])+alloc+1;
}
/* Return the pointer of the actual SDS allocation (normally SDS strings
* are referenced by the start of the string buffer). */
/**
*
* @param s
* @return sds起始指针
*/
void *sdsAllocPtr(sds s) {
return (void*) (s-sdsHdrSize(s[-1]));
}
/* Increment the sds length and decrements the left free space at the
* end of the string according to 'incr'. Also set the null term
* in the new end of the string.
*
* This function is used in order to fix the string length after the
* user calls sdsMakeRoomFor(), writes something after the end of
* the current string, and finally needs to set the new length.
*
* Note: it is possible to use a negative increment in order to
* right-trim the string.
*
* Usage example:
*
* Using sdsIncrLen() and sdsMakeRoomFor() it is possible to mount the
* following schema, to cat bytes coming from the kernel to the end of an
* sds string without copying into an intermediate buffer:
*
* oldlen = sdslen(s);
* s = sdsMakeRoomFor(s, BUFFER_SIZE);
* nread = read(fd, s+oldlen, BUFFER_SIZE);
* ... check for nread <= 0 and handle it ...
* sdsIncrLen(s, nread);
*/
/**
* 将已使用长度,偏移到incr,若incr是负数,则向左截断.
* @param s
* @param incr
*/
void sdsIncrLen(sds s, ssize_t incr) {
unsigned char flags = s[-1];
size_t len;
switch(flags&SDS_TYPE_MASK) {
case SDS_TYPE_5: {
unsigned char *fp = ((unsigned char*)s)-1;
unsigned char oldlen = SDS_TYPE_5_LEN(flags);
assert((incr > 0 && oldlen+incr < 32) || (incr < 0 && oldlen >= (unsigned int)(-incr)));
*fp = SDS_TYPE_5 | ((oldlen+incr) << SDS_TYPE_BITS);
len = oldlen+incr;
break;
}
case SDS_TYPE_8: {
SDS_HDR_VAR(8,s);
assert((incr >= 0 && sh->alloc-sh->len >= incr) || (incr < 0 && sh->len >= (unsigned int)(-incr)));
len = (sh->len += incr);
break;
}
case SDS_TYPE_16: {
SDS_HDR_VAR(16,s);
assert((incr >= 0 && sh->alloc-sh->len >= incr) || (incr < 0 && sh->len >= (unsigned int)(-incr)));
len = (sh->len += incr);
break;
}
case SDS_TYPE_32: {
SDS_HDR_VAR(32,s);
assert((incr >= 0 && sh->alloc-sh->len >= (unsigned int)incr) || (incr < 0 && sh->len >= (unsigned int)(-incr)));
len = (sh->len += incr);
break;
}
case SDS_TYPE_64: {
SDS_HDR_VAR(64,s);
assert((incr >= 0 && sh->alloc-sh->len >= (uint64_t)incr) || (incr < 0 && sh->len >= (uint64_t)(-incr)));
len = (sh->len += incr);
break;
}
default: len = 0; /* Just to avoid compilation warnings. */
}
//标记偏移最后一位为空字符串
s[len] = '\0';
}
/* Grow the sds to have the specified length. Bytes that were not part of
* the original length of the sds will be set to zero.
*
* if the specified length is smaller than the current length, no operation
* is performed. */
/**
* 将 sds 扩充至指定长度,未使用的空间以 0 字节填充。
* * 复杂度:
* T = O(N)
* @param s
* @param len
* @return
*/
sds sdsgrowzero(sds s, size_t len) {
size_t curlen = sdslen(s);
if (len <= curlen) return s;
s = sdsMakeRoomFor(s,len-curlen);
if (s == NULL) return NULL;
/* Make sure added region doesn't contain garbage */
//尚未使用的空间,填充0
memset(s+curlen,0,(len-curlen+1)); /* also set trailing \0 byte */
//设置新的使用空间的长度
sdssetlen(s, len);
return s;
}
/* Append the specified binary-safe string pointed by 't' of 'len' bytes to the
* end of the specified sds string 's'.
*
* After the call, the passed sds string is no longer valid and all the
* references must be substituted with the new pointer returned by the call. */
/**
* 将len长度的t字符串,追加到s
* @param s
* @param t
* @param len
* @return
*/
sds sdscatlen(sds s, const void *t, size_t len) {
size_t curlen = sdslen(s);
//扩容
s = sdsMakeRoomFor(s,len);
// 内存不足,直接返回
if (s == NULL) return NULL;
//将t追加到sds字符串末尾
memcpy(s+curlen, t, len);
//更改长度
sdssetlen(s, curlen+len);
//添加新结尾符号
s[curlen+len] = '\0';
//返回sds
return s;
}
/* Append the specified null termianted C string to the sds string 's'.
*
* After the call, the passed sds string is no longer valid and all the
* references must be substituted with the new pointer returned by the call. */
/**
* 将t追加到s
* @param s
* @param t
* @return
*/
sds sdscat(sds s, const char *t) {
return sdscatlen(s, t, strlen(t));
}
/* Append the specified sds 't' to the existing sds 's'.
*
* After the call, the modified sds string is no longer valid and all the
* references must be substituted with the new pointer returned by the call. */
/**
* 将另一个sds追加到s
* @param s
* @param t
* @return
*/
sds sdscatsds(sds s, const sds t) {
return sdscatlen(s, t, sdslen(t));
}
/* Destructively modify the sds string 's' to hold the specified binary
* safe string pointed by 't' of length 'len' bytes. */
/**
* 将len长度的t复制到s中,会覆盖s内容,内存不足则重新分配
* *
* 复杂度
* T = O(N)
* @param s
* @param t
* @param len
* @return
*/
sds sdscpylen(sds s, const char *t, size_t len) {
//分配的长度<len,则扩容
if (sdsalloc(s) < len) {
s = sdsMakeRoomFor(s,len-sdslen(s));
//内存不足,则返回NULL
if (s == NULL) return NULL;
}
//复制内容
memcpy(s, t, len);
s[len] = '\0';
//更新属性
sdssetlen(s, len);
return s;
}
/**
* 将复制到s中,会覆盖s内容,内存不足则重新分配
* @param s
* @param t
* @return
*/
/* Like sdscpylen() but 't' must be a null-termined string so that the length
* of the string is obtained with strlen(). */
sds sdscpy(sds s, const char *t) {
return sdscpylen(s, t, strlen(t));
}
/* Helper for sdscatlonglong() doing the actual number -> string
* conversion. 's' must point to a string with room for at least
* SDS_LLSTR_SIZE bytes.
*
* The function returns the length of the null-terminated string
* representation stored at 's'. */
#define SDS_LLSTR_SIZE 21
/**
* 将 ll 类型的value转成字符串,并存入s
* @param s
* @param value
* @return
*/
int sdsll2str(char *s, long long value) {
char *p, aux;
unsigned long long v;
size_t l;
/* Generate the string representation, this method produces
* an reversed string. */
v = (value < 0) ? -value : value;
p = s;
do {
*p++ = '0'+(v%10);
v /= 10;
} while(v);
if (value < 0) *p++ = '-';
/* Compute length and add null term. */
l = p-s;
*p = '\0';
/* Reverse the string. */
p--;
while(s < p) {
aux = *s;
*s = *p;
*p = aux;
s++;
p--;
}
return l;
}
/* Identical sdsll2str(), but for unsigned long long type. */
/**
* 将 ull 类型的value转成字符串,并存入s
* @param s
* @param v
* @return
*/
int sdsull2str(char *s, unsigned long long v) {
char *p, aux;
size_t l;
/* Generate the string representation, this method produces
* an reversed string. */
p = s;
do {
*p++ = '0'+(v%10);
v /= 10;
} while(v);
/* Compute length and add null term. */
l = p-s;
*p = '\0';
/* Reverse the string. */
p--;
while(s < p) {
aux = *s;
*s = *p;
*p = aux;
s++;
p--;
}
return l;
}
/* Create an sds string from a long long value. It is much faster than:
*
* sdscatprintf(sdsempty(),"%lld\n", value);
*/
/**
* 将long long类型value转成string,然后根据value新建sds
* @param value
* @return
*/
sds sdsfromlonglong(long long value) {
char buf[SDS_LLSTR_SIZE];
int len = sdsll2str(buf,value);
return sdsnewlen(buf,len);
}
/* Like sdscatprintf() but gets va_list instead of being variadic. */
sds sdscatvprintf(sds s, const char *fmt, va_list ap) {
va_list cpy;
char staticbuf[1024], *buf = staticbuf, *t;
size_t buflen = strlen(fmt)*2;
/* We try to start using a static buffer for speed.
* If not possible we revert to heap allocation. */
if (buflen > sizeof(staticbuf)) {
buf = s_malloc(buflen);
if (buf == NULL) return NULL;
} else {
buflen = sizeof(staticbuf);
}
/* Try with buffers two times bigger every time we fail to
* fit the string in the current buffer size. */
while(1) {
buf[buflen-2] = '\0';
va_copy(cpy,ap);
vsnprintf(buf, buflen, fmt, cpy);
va_end(cpy);
if (buf[buflen-2] != '\0') {
if (buf != staticbuf) s_free(buf);
buflen *= 2;
buf = s_malloc(buflen);
if (buf == NULL) return NULL;
continue;
}
break;
}
/* Finally concat the obtained string to the SDS string and return it. */
t = sdscat(s, buf);
if (buf != staticbuf) s_free(buf);
return t;
}
/* Append to the sds string 's' a string obtained using printf-alike format
* specifier.
*
* After the call, the modified sds string is no longer valid and all the
* references must be substituted with the new pointer returned by the call.
*
* Example:
*
* s = sdsnew("Sum is: ");
* s = sdscatprintf(s,"%d+%d = %d",a,b,a+b).
*
* Often you need to create a string from scratch with the printf-alike
* format. When this is the need, just use sdsempty() as the target string:
*
* s = sdscatprintf(sdsempty(), "... your format ...", args);
*/
sds sdscatprintf(sds s, const char *fmt, ...) {
va_list ap;
char *t;
va_start(ap, fmt);
t = sdscatvprintf(s,fmt,ap);
va_end(ap);
return t;
}
/* This function is similar to sdscatprintf, but much faster as it does
* not rely on sprintf() family functions implemented by the libc that
* are often very slow. Moreover directly handling the sds string as
* new data is concatenated provides a performance improvement.
*
* However this function only handles an incompatible subset of printf-alike
* format specifiers:
*
* %s - C String
* %S - SDS string
* %i - signed int
* %I - 64 bit signed integer (long long, int64_t)
* %u - unsigned int
* %U - 64 bit unsigned integer (unsigned long long, uint64_t)
* %% - Verbatim "%" character.
*/
sds sdscatfmt(sds s, char const *fmt, ...) {
size_t initlen = sdslen(s);
const char *f = fmt;
long i;
va_list ap;
/* To avoid continuous reallocations, let's start with a buffer that
* can hold at least two times the format string itself. It's not the
* best heuristic but seems to work in practice. */
s = sdsMakeRoomFor(s, initlen + strlen(fmt)*2);
va_start(ap,fmt);
f = fmt; /* Next format specifier byte to process. */
i = initlen; /* Position of the next byte to write to dest str. */
while(*f) {
char next, *str;
size_t l;
long long num;
unsigned long long unum;
/* Make sure there is always space for at least 1 char. */
if (sdsavail(s)==0) {
s = sdsMakeRoomFor(s,1);
}
switch(*f) {
case '%':
next = *(f+1);
f++;
switch(next) {
case 's':
case 'S':
str = va_arg(ap,char*);
l = (next == 's') ? strlen(str) : sdslen(str);
if (sdsavail(s) < l) {
s = sdsMakeRoomFor(s,l);
}
memcpy(s+i,str,l);
sdsinclen(s,l);
i += l;
break;
case 'i':
case 'I':
if (next == 'i')
num = va_arg(ap,int);
else
num = va_arg(ap,long long);
{
char buf[SDS_LLSTR_SIZE];
l = sdsll2str(buf,num);
if (sdsavail(s) < l) {
s = sdsMakeRoomFor(s,l);
}
memcpy(s+i,buf,l);
sdsinclen(s,l);
i += l;
}
break;
case 'u':
case 'U':
if (next == 'u')
unum = va_arg(ap,unsigned int);
else
unum = va_arg(ap,unsigned long long);
{
char buf[SDS_LLSTR_SIZE];
l = sdsull2str(buf,unum);
if (sdsavail(s) < l) {
s = sdsMakeRoomFor(s,l);
}
memcpy(s+i,buf,l);
sdsinclen(s,l);
i += l;
}
break;
default: /* Handle %% and generally %<unknown>. */
s[i++] = next;
sdsinclen(s,1);
break;
}
break;
default:
s[i++] = *f;
sdsinclen(s,1);
break;
}
f++;
}
va_end(ap);
/* Add null-term */
s[i] = '\0';
return s;
}
/* Remove the part of the string from left and from right composed just of
* contiguous characters found in 'cset', that is a null terminted C string.
*
* After the call, the modified sds string is no longer valid and all the
* references must be substituted with the new pointer returned by the call.
*
* Example:
*
* s = sdsnew("AA...AA.a.aa.aHelloWorld :::");
* s = sdstrim(s,"Aa. :");
* printf("%s\n", s);
*
* Output will be just "HelloWorld".
*/
/**
* 对sds左右二端修剪,清除cset指定的所有字符
* * 复杂度:
* T = O(M*N),M 为 SDS 长度, N 为 cset 长度。
* @param s
* @param cset
* @return
*/
sds sdstrim(sds s, const char *cset) {
char *start, *end, *sp, *ep;
size_t len;
sp = start = s;
ep = end = s+sdslen(s)-1;
//分别用sp和ep标记剩余字符非cset的指针
while(sp <= end && strchr(cset, *sp)) sp++;
while(ep > sp && strchr(cset, *ep)) ep--;
//重新计算去除cset后剩余长度
len = (sp > ep) ? 0 : ((ep-sp)+1);
//将sp~ep中间的字符,移动到sds
if (s != sp) memmove(s, sp, len);
//标记最后一个字符为空字符串
s[len] = '\0';
//更新长度
sdssetlen(s,len);
return s;
}
/* Turn the string into a smaller (or equal) string containing only the
* substring specified by the 'start' and 'end' indexes.
*
* start and end can be negative, where -1 means the last character of the
* string, -2 the penultimate character, and so forth.
*
* The interval is inclusive, so the start and end characters will be part
* of the resulting string.
*
* The string is modified in-place.
*
* Example:
*
* s = sdsnew("Hello World");
* sdsrange(s,1,-1); => "ello World"
*/
/**
* 截断start~end字符
* * 复杂度
* T = O(N)
* @param s
* @param start
* @param end
*/
void sdsrange(sds s, ssize_t start, ssize_t end) {
size_t newlen, len = sdslen(s);
if (len == 0) return;
if (start < 0) {
start = len+start;
if (start < 0) start = 0;
}
if (end < 0) {
end = len+end;
if (end < 0) end = 0;
}
newlen = (start > end) ? 0 : (end-start)+1;
if (newlen != 0) {
if (start >= (ssize_t)len) {
newlen = 0;
} else if (end >= (ssize_t)len) {
end = len-1;
newlen = (start > end) ? 0 : (end-start)+1;
}
} else {
start = 0;
}
if (start && newlen) memmove(s, s+start, newlen);
s[newlen] = 0;
sdssetlen(s,newlen);
}
/* Apply tolower() to every character of the sds string 's'. */
/**
* 将sds转成小写
* @param s
*/
void sdstolower(sds s) {
size_t len = sdslen(s), j;
for (j = 0; j < len; j++) s[j] = tolower(s[j]);
}
/* Apply toupper() to every character of the sds string 's'. */
/**
*将sds转成大写
* @param s
*/
void sdstoupper(sds s) {
size_t len = sdslen(s), j;
for (j = 0; j < len; j++) s[j] = toupper(s[j]);
}
/* Compare two sds strings s1 and s2 with memcmp().
*
* Return value:
*
* positive if s1 > s2.
* negative if s1 < s2.
* 0 if s1 and s2 are exactly the same binary string.
*
* If two strings share exactly the same prefix, but one of the two has
* additional characters, the longer string is considered to be greater than
* the smaller one. */
/**
* 比较
* * T = O(N)
* @param s1
* @param s2
* @return
*/
int sdscmp(const sds s1, const sds s2) {
size_t l1, l2, minlen;
int cmp;
l1 = sdslen(s1);
l2 = sdslen(s2);
minlen = (l1 < l2) ? l1 : l2;
cmp = memcmp(s1,s2,minlen);
if (cmp == 0) return l1>l2? 1: (l1<l2? -1: 0);
return cmp;
}
链表
adlist.h
链表数据结构
typedef struct list {
//头节点
listNode *head;
//尾节点
listNode *tail;
//复制节点保存值
void *(*dup)(void *ptr);
//释放节点保存值
void (*free)(void *ptr);
//对比节点和另外一个输入值是否相等
int (*match)(void *ptr, void *key);
//节点数
unsigned long len;
} list;
链表节点结构
typedef struct listNode {
//前置节点
struct listNode *prev;
//后置节点
struct listNode *next;
//值
void *value;
} listNode;
字典
数据结构如下
dict.h
字典表
typedef struct dict {
//类型
dictType *type;
//私有数据
void *privdata;
//hash表,ht[1]只会对ht[0]进行rehash时使用
dictht ht[2];
//记录rehash进度,rehash不在进行时,为-1
long rehashidx; /* rehashing not in progress if rehashidx == -1 */
unsigned long iterators; /* number of iterators currently running */
} dict;
字典类型
typedef struct dictType {
//算hash值得
uint64_t (*hashFunction)(const void *key);
//复制键
void *(*keyDup)(void *privdata, const void *key);
//复制value
void *(*valDup)(void *privdata, const void *obj);
//比较键
int (*keyCompare)(void *privdata, const void *key1, const void *key2);
//销毁键
void (*keyDestructor)(void *privdata, void *key);
//销毁val
void (*valDestructor)(void *privdata, void *obj);
} dictType;
字段hash表
typedef struct dictht {
//hash表数组,存键值对
dictEntry **table;
//table总容量
unsigned long size;
//hasn表大小掩码,计算索引值,总是为size-1
unsigned long sizemask;
//键值对数量,包含因为冲突形成链表的键值对
unsigned long used;
} dictht;
键值对
typedef struct dictEntry {
//键
void *key;
//值
union {
void *val;
uint64_t u64;
int64_t s64;
double d;
} v;
//下个hash节点,解决hash碰撞
struct dictEntry *next;
} dictEntry;
如何解决hash冲突的?
采用链地址法.采用头插法.
点击这里可以参考java实现hash链地址法
也点击这里参考CHM源码,包括扩容机制
(How)如何扩容和收缩?
- 为ht[1]分配空间
若扩展,则ht[0].used * 2,然后在used~uses*2范围内取出>=used的2^n
若缩减,找出>=used的2^n - 将ht[0]所有键值对重新计算hashCode和索引,然后存ht[1]中
- 释放ht[0],变更ht[1]为ht[0],然后新建空白ht[1]
这里我们可以看到每次rehash后,扩容大小绝对为2^N,这样有啥意义了?大家有没有发现计算索引的sizemask,这个值为容量-1,而容量为2的N次幂,这就保证了sizemask二进制表示为1111111,然后每次hash取模时,都能均匀分布.
(When)何时扩容和收缩?
扩容时机:
- 服务器没有执行BGSAVE或BGREWRITEAOF命令,且hash表负载因子>=1
- 服务器正在执行BGSAVE或BGREWRITEAOF命令,且hash表负载因子>=5
负载因子=ht[0].used/ht[0].size
可以发现当执行BGSAVE或BGREWRITEAOF命令时,负载因子变得很大,这是因为刷入磁盘需要占用大量磁盘IO,为了减少IO争抢,而扩展负载因子.
收缩时机:
负载因子<0.1
当数据量无比庞大,如何扩容?
当键值对有成百上千时,如果一次性rehash,这对服务器来说可能暂时降低吞吐,那么这里就要谈谈rehash的渐进式策略.也就是rehash分多次完成.
步骤如下:
- rehash时,设置rehashidx=0,表示开始rehash了
- 然后在每次对字典CURD时,都会将ht[0].table.dictEntry[rehashidx]保存ht[1]中,然后rehashidx+=1
- rehash完成,rehashidx=-1
这样的优点,将性能消耗平摊到每次CRUD上,而不是集中消耗性能.
针对R时,会先去ht[0]查,查不到接着ht[1]查.
在C时,一律保存ht[1]中,
这保证了rehash期间ht[0]不作任何操作,最终数据都会到ht[1]中.
跳跃表
场景
有序集合
集群节点内部数据
server.h
//跳跃表节点
typedef struct zskiplistNode {
//元素
sds ele;
//分值,节点按分值从小到大排列
double score;
//后退,指向当前节点的前一个节点,主要用于表尾向表头遍历使用
struct zskiplistNode *backward;
//节点层
struct zskiplistLevel {
//前进指针,用于访问位于表尾方向的其他节点,主要用于表头向表尾遍历使用
struct zskiplistNode *forward;
//跨度,记录前进指针所指向节点和当前节点的距离
unsigned long span;
} level[];
} zskiplistNode;
//跳跃表列表
typedef struct zskiplist {
//头,尾节点
struct zskiplistNode *header, *tail;
//节点数(不包含头节点)
unsigned long length;
//所有节点最大的层数(不包含头节点)
int level;
} zskiplist;
整数集合
intset.h
//整数数组
typedef struct intset{
//编码
uint32_t encoding;
//元素数量
uint32_t length;
//保存元素的数组,唯一,有序排列
//具体存储的位数与编码有关,若编码是int16,则16位存.每个数组值都是16位存储
int8_t contents[];
} intset;
升级or降级
如果之前存了很多16位数据,接着插入了一个32位数据,就需要将之前16位数据升级为64位.
升级的步骤如下
- 分配所有元素占用32位所需空间
- 接着移动之前16位最后一个数据到重新计算32位的索引位置.
- 移动完成后,最后新加入的32位数据插入最后一个位.
优点:
可存储任意类型数据
当存储高位数时,才会考虑升级,节省内存
不存在降级
压缩列表
使用场景:
List
Hash