There are various kinds of identifiers and handles used across FiveM. A few of them need special attention and are mentioned below.
The server ID, canonically known as the NetID, is a transient handle that is usually seen in server scripts as
To convert between a player index on the client and a server index, use GET_PLAYER_FROM_SERVER_ID and GET_PLAYER_SERVER_ID on the client.
The player index is used on the client to refer to a player. It is usually a number from 0 to 128 (inclusive), and is used for game natives as well as some
CFX natives on the client.
With state awareness enabled, player indices are local to a specific client, as with entity handles.
Note that a player existent on the server does not necessarily exist on the client as well: GET_PLAYER_FROM_SERVER_ID can return
-1 if this is the case.
A player will usually have a ped representing the physical entity making up the player's character. This is a standard network object with a network ID.
To get a player's ped, use GET_PLAYER_PED, and to get the player belonging to a ped handle, use NETWORK_GET_PLAYER_INDEX_FROM_PED.
An entity handle (Entity, Ped, etc) refers to the internal game representation of the entity, and is local to a specific client. This means you can not expect the same entity to have the same handle on different clients.
Entity handles do not tend to be reused for different entities, however the same entity - if recreated locally or remotely - may get a different handle.
To convert from an entity handle to a network ID and back, use NETWORK_GET_NETWORK_ID_FROM_ENTITY and NETWORK_GET_ENTITY_FROM_NETWORK_ID (or the equivalent aliases such as PED_TO_NET, NET_TO_VEH, etc.).
An entity network ID is a 16-bit integer representing a given entity existing at a point in time. These can and will be reused, but during the lifetime of an entity, the network ID will not change.
If you need to refer to an entity across machines (clients, or the server), you should use its network ID.
Note that just like with players, a network ID is not necessarily in scope on a client: verify its existence using NETWORK_DOES_ENTITY_EXIST_WITH_NETWORK_ID before using it.