badger

README

Badger Store

Tables

Since BadgerDB has a single namespace, we use key prefixes to designate separate namespaces. These namespaces act as tables or indexes. Below are the namespaces that are used internally, along with their single-byte prefixes:

Prefix	Namespace	Description
0x00	Idents	Holds covering index of (ID, name) for each ident
0x01	IdentIDByName	Mapping of Ident name to Ident ID
0x02	EAVT	(Entity, Attribute) -> (Value, Tx)
0x03	AEVT	(Attribute, Entity) -> (Value, Tx)
0x04	AVET	(Attribute, Value) -> (Entity, Tx)
0x05	VAET	(Value, Attribute) -> (Entity, Tx)

Ident Storage

Since Badger keeps keys in LSM trees, and there should always be a small number of idents, we keep all idents in a covering index whose key encodes both the ID and name of each ident. This should allow us to look up any ident by name extremely quickly. Additionally, we keep a separate index of (ID, Name), but in this case, the name is stored as a value and will not be kept in memory. Thus, ident lookups by ID will be slower, but this will not be a common operation.

Documentation

Index

• func New(db *badger.DB) (*badgerStore, error)

Functions

func New

func New(db *badger.DB) (*badgerStore, error)