In PuzzleDB, records, schemas, and indices are all represented as key-value pairs. This section explains the format of the key-value data used in PuzzleDB.
The key-value store is a collection of key-value records, where each record is a key-value pair, consisting of a header as the key. The key-value store supports the following categories of key-value records:
| Category | Key Order | Value | |||
|---|---|---|---|---|---|
0 |
1 |
2 |
3 |
||
Database |
Header (D) |
Database |
- |
- |
CBOR (Options) |
Collection |
Header © |
Database |
Collection |
- |
CBOR (Schema) |
Object |
Header (O) |
Database |
Collection |
Element |
CBOR (Object) |
Index |
Header (I) |
Database |
Collection |
Element |
Tuple (Key) |
The key header is a 2-byte header that is prepended to every key in the key-value store. The key header is reserved as follows:
| Field Name | Size (bits) | Description | Example Value |
|---|---|---|---|
Key category |
8 |
The record key type |
D:Database C:Collection O:Document I:Index |
Version |
4 |
The version number |
0:reserved 1-7 |
Value type |
4 |
The record value type |
0:reserved 1:CBOR 1:PRIMARY 2:SECONDARY |
The key header begins with a 1-byte identifier for the key type, enabling key type-based searching. Duplication is tolerated because a value type is reserved for each key type.