Redis underlying data structure -- listpack

listpack Introduce

Redis stay 5.0 A new data structure called listpack, The purpose is to replace compressed lists , Its biggest feature is listpack Each node in no longer contains the length of the previous node , Each node in the compressed list needs to save the length field of the previous node , There will be a hidden danger of chain update .

listpack The structure design

listpack Many excellent designs with compressed lists , such as Or use a contiguous memory space to store data compactly , And to save memory overhead ,listpack Nodes use different encoding methods to save data of different sizes .

listpack  There are four parameters 
 Total bytes    Element quantity   Node node     End tag 

listpack entry  The node structure 
encoding  Define the encoding type of the element , Will encode integers and strings of different lengths 
data  Data actually stored 
len,encoding+data The total length of 

listpack There is no field in the compressed list to record the length of the previous node ,listpack Only record the length of the current node , When we turn to listpack When adding a new element , It will not affect the change of the length field of other nodes , Thus, the chain update problem of compressed list is avoided .