mysql存储map数据结构

Go map实现原理 - 恋恋美食的个人空间 - OSCHINA - 中文开源技术交流社区 https://my.oschina.net/renhc/blog/2208417

// A header for a Go map.

type hmap struct {

// Note: the format of the hmap is also encoded in cmd/compile/internal/gc/reflect.go.

// Make sure this stays in sync with the compiler's definition.

count int // # live cells == size of map. Must be first (used by len() builtin)

flags uint8

B uint8 // log_2 of # of buckets (can hold up to loadFactor * 2^B items)

noverflow uint16 // approximate number of overflow buckets; see incrnoverflow for details

hash0 uint32 // hash seed

buckets unsafe.Pointer // array of 2^B Buckets. may be nil if count==0.

oldbuckets unsafe.Pointer // previous bucket array of half the size, non-nil only when growing

nevacuate uintptr // progress counter for evacuation (buckets less than this have been evacuated)

extra *mapextra // optional fields

}

Golang的map使用哈希表作为底层实现，一个哈希表里可以有多个哈希表节点，也即bucket，而每个bucket就保存了map中的一个或一组键值对。

map数据结构由runtime/map.go:hmap定义:

type hmapstruct{

countint//当前保存的元素个数

...

B uint8

...

bucketsunsafe.Pointer// bucket数组指针，数组的大小为2^B

...

}

下图展示一个拥有4个bucket的map：

本例中, hmap.B=2，而hmap.buckets长度是2^B为4. 元素经过哈希运算后会落到某个bucket中进行存储。查找过程类似。

bucket很多时候被翻译为桶，所谓的哈希桶实际上就是bucket。

2. bucket数据结构

bucket数据结构由runtime/map.go:bmap定义：

type bmapstruct{

tophash [8]uint8//存储哈希值的高8位

databyte[1]//key value数据:key/key/key/.../value/value/value...

overflow *bmap//溢出bucket的地址

}

每个bucket可以存储8个键值对。

tophash是个长度为8的数组，哈希值相同的键(准确的说是哈希值低位相同的键)存入当前bucket时会将哈希值的高位存储在该数组中，以方便后续匹配。

data区存放的是key-value数据，存放顺序是key/key/key/…value/value/value，如此存放是为了节省字节对齐带来的空间浪费。

overflow 指针指向的是下一个bucket，据此将所有冲突的键连接起来。

注意：上述中data和overflow并不是在结构体中显示定义的，而是直接通过指针运算进行访问的。

下图展示bucket存放8个key-value对：

package runtime

// This file contains the implementation of Go's map type.

// A map is just a hash table. The data is arranged

// into an array of buckets. Each bucket contains up to

// 8 key/elem pairs. The low-order bits of the hash are

// used to select a bucket. Each bucket contains a few

// high-order bits of each hash to distinguish the entries

// within a single bucket.

// If more than 8 keys hash to a bucket, we chain on

// extra buckets.

// When the hashtable grows, we allocate a new array

// of buckets twice as big. Buckets are incrementally

// copied from the old bucket array to the new bucket array.

// Map iterators walk through the array of buckets and

// return the keys in walk order (bucket #, then overflow

// chain order, then bucket index). To maintain iteration

// semantics, we never move keys within their bucket (if

// we did, keys might be returned 0 or 2 times). When

// growing the table, iterators remain iterating through the

// old table and must check the new table if the bucket

// they are iterating through has been moved ("evacuated")

// to the new table.

// Picking loadFactor: too large and we have lots of overflow

// buckets, too small and we waste a lot of space. I wrote

// a simple program to check some stats for different loads:

// (64-bit, 8 byte keys and elems)

// loadFactor %overflow bytes/entry hitprobe missprobe

// 4.00 2.13 20.77 3.00 4.00

// 4.50 4.05 17.30 3.25 4.50

// 5.00 6.85 14.77 3.50 5.00

// 5.50 10.55 12.94 3.75 5.50

// 6.00 15.27 11.67 4.00 6.00

// 6.50 20.90 10.79 4.25 6.50

// 7.00 27.14 10.15 4.50 7.00

// 7.50 34.03 9.73 4.75 7.50

// 8.00 41.10 9.40 5.00 8.00

// %overflow = percentage of buckets which have an overflow bucket

// bytes/entry = overhead bytes used per key/elem pair

// hitprobe = # of entries to check when looking up a present key

// missprobe = # of entries to check when looking up an absent key

// Keep in mind this data is for maximally loaded tables, i.e. just

// before the table grows. Typical tables will be somewhat less loaded.

import (

"runtime/internal/atomic"

"runtime/internal/math"

"runtime/internal/sys"

"unsafe"

)

const (

// Maximum number of key/elem pairs a bucket can hold.

bucketCntBits = 3

bucketCnt = 1 << bucketCntBits

// Maximum average load of a bucket that triggers growth is 6.5.

// Represent as loadFactorNum/loadFactorDen, to allow integer math.

loadFactorNum = 13

loadFactorDen = 2

// Maximum key or elem size to keep inline (instead of mallocing per element).

// Must fit in a uint8.

// Fast versions cannot handle big elems - the cutoff size for

// fast versions in cmd/compile/internal/gc/walk.go must be at most this elem.

maxKeySize = 128

maxElemSize = 128

// data offset should be the size of the bmap struct, but needs to be

// aligned correctly. For amd64p32 this means 64-bit alignment

// even though pointers are 32 bit.

dataOffset = unsafe.Offsetof(struct {

b bmap

v int64

}{}.v)

// Possible tophash values. We reserve a few possibilities for special marks.

// Each bucket (including its overflow buckets, if any) will have either all or none of its

// entries in the evacuated* states (except during the evacuate() method, which only happens

// during map writes and thus no one else can observe the map during that time).

emptyRest = 0 // this cell is empty, and there are no more non-empty cells at higher indexes or overflows.

emptyOne = 1 // this cell is empty

evacuatedX = 2 // key/elem is valid. Entry has been evacuated to first half of larger table.

evacuatedY = 3 // same as above, but evacuated to second half of larger table.

evacuatedEmpty = 4 // cell is empty, bucket is evacuated.

minTopHash = 5 // minimum tophash for a normal filled cell.

// flags

iterator = 1 // there may be an iterator using buckets

oldIterator = 2 // there may be an iterator using oldbuckets

hashWriting = 4 // a goroutine is writing to the map

sameSizeGrow = 8 // the current map growth is to a new map of the same size

// sentinel bucket ID for iterator checks

noCheck = 1<

)

3. 哈希冲突

当有两个或以上数量的键被哈希到了同一个bucket时，我们称这些键发生了冲突。Go使用链地址法来解决键冲突。由于每个bucket可以存放8个键值对，所以同一个bucket存放超过8个键值对时就会再创建一个键值对，用类似链表的方式将bucket连接起来。

下图展示产生冲突后的map：

bucket数据结构指示下一个bucket的指针称为overflow bucket，意为当前bucket盛不下而溢出的部分。事实上哈希冲突并不是好事情，它降低了存取效率，好的哈希算法可以保证哈希值的随机性，但冲突过多也是要控制的，后面会再详细介绍。

4. 负载因子

负载因子用于衡量一个哈希表冲突情况，公式为：

负载因子 = 键数量/bucket数量

例如，对于一个bucket数量为4，包含4个键值对的哈希表来说，这个哈希表的负载因子为1.

哈希表需要将负载因子控制在合适的大小，超过其阀值需要进行rehash，也即键值对重新组织：

哈希因子过小，说明空间利用率低

哈希因子过大，说明冲突严重，存取效率低

每个哈希表的实现对负载因子容忍程度不同，比如Redis实现中负载因子大于1时就会触发rehash，而Go则在在负载因子达到6.5时才会触发rehash，因为Redis的每个bucket只能存1个键值对，而Go的bucket可能存8个键值对，所以Go可以容忍更高的负载因子。

5. 渐进式扩容

5.1 扩容的前提条件

为了保证访问效率，当新元素将要添加进map时，都会检查是否需要扩容，扩容实际上是以空间换时间的手段。

触发扩容的条件有二个：

1. 负载因子 > 6.5时，也即平均每个bucket存储的键值对达到6.5个。

2. overflow数量 > 2^15时，也即overflow数量超过32768时。

5.2 增量扩容

当负载因子过大时，就新建一个bucket，新的bucket长度是原来的2倍，然后旧bucket数据搬迁到新的bucket。

考虑到如果map存储了数以亿计的key-value，一次性搬迁将会造成比较大的延时，Go采用逐步搬迁策略，即每次访问map时都会触发一次搬迁，每次搬迁2个键值对。

下图展示了包含一个bucket满载的map(为了描述方便，图中bucket省略了value区域):

当前map存储了7个键值对，只有1个bucket。此地负载因子为7。再次插入数据时将会触发扩容操作，扩容之后再将新插入键写入新的bucket。

当第8个键值对插入时，将会触发扩容，扩容后示意图如下：

hmap数据结构中oldbuckets成员指身原bucket，而buckets指向了新申请的bucket。新的键值对被插入新的bucket中。后续对map的访问操作会触发迁移，将oldbuckets中的键值对逐步的搬迁过来。当oldbuckets中的键值对全部搬迁完毕后，删除oldbuckets。

搬迁完成后的示意图如下：

数据搬迁过程中原bucket中的键值对将存在于新bucket的前面，新插入的键值对将存在于新bucket的后面。实际搬迁过程中比较复杂，将在后续源码分析中详细介绍。

5.3 等量扩容

所谓等量扩容，实际上并不是扩大容量，buckets数量不变，重新做一遍类似增量扩容的搬迁动作，把松散的键值对重新排列一次，以使bucket的使用率更高，进而保证更快的存取。在极端场景下，比如不断地增删，而键值对正好集中在一小部分的bucket，这样会造成overflow的bucket数量增多，但负载因子又不高，从而无法执行增量搬迁的情况，如下图所示：