The internal coding of redis basic data structures

preface

• The five data structures of Redis are string, Hash, list, set, and zset. There are actually multiple implementations of their internal code.

• Redis can view the internal encoding mode using the object Encoding command.

parsing

• Three internal encodings of String

  • Int: a long integer of 8 bytes.
  • Embstr: contains a string of 39 bytes or less.
  • Raw: The string contains more than 39 bytes.

• There are two internal encodings of hash:

  • Ziplist: When the number of hash elements is smaller than hash-max-ziplist-entries (512 by default) and the element size is smaller than hash-max-ziplist-value(64 bytes by default). Redis uses Ziplist as an internal implementation of the hash.
  • Hashtable (hashtable) : When the hash does not satisfy the Ziplist implementation, the hashtable is used as the internal hash implementation. In this case, the ziplist read/write performance deteriorates, while the time complexity of hashTable is O(1).
  • Note: Redis will always use hashtable as its internal implementation after ziplist is converted to Hashtable, even if the contents of the subsequent cache are completely replaced.

• There are three internal encodings of list:

  • Ziplist: Redis uses ziplist as an internal implementation of lists when the number of elements in a list is less than hash-max-ziplist-entries and the size of the elements is less than hash-max-ziplist-value.
  • Linkedlist (linkedlist) : when a list does not satisfy a ziplist implementation, the linkedlist is used as an internal implementation.
  • Quicklist: This data structure is new since Redis 3.2 and combines ziplist+ LinkedList.

- Note that lists, like hash, can only have ziplist turned into linkedList, not linkedList turned into Ziplist.Copy the code

• Two internal encodings of a set:

  • Intset: Redis uses intset as an implementation of collections where all elements are integers and the number of entries is less than set-max-intset-entries (default: 512) to reduce memory usage.
  • Hashtable: Redis uses hashtable as the collection implementation when the collection does not satisfy intSet implementation.

• Two internal encodings of Zset:

  • Ziplist: Zset-max-ziplist-entries are ordered when the number of elements is less than or equal to zset-max-ziplist-value (default: 128) and each element is less than zset-max-ziplist-value (default: 64 bytes). Redis uses Ziplist as an internal collection implementation, which reduces memory usage.
  • Skiplist: Use skiplist as an internal implementation when ordered collections do not satisfy Ziplist, because ziplist reads and writes are less efficient.

The last

• Redis also has several advanced data structures, including Bitmaps. It is similar to the String data structure, but instead of a string, it stores bits, so a string structure can store 512 megabytes, and a bitmap can store 2^32 bits. In addition, distributed BloomFilter is based on this.
• Learn with an open mind and make progress together