Reference fields are pointers to other Documents. They are indexed automatically and can be queried with filter.

To ensure type safety, use and store only references returned by the primitive queries, like insert, range or filter. Numerical instances like Num or Integral are provided to support generic database programs that take the responsibility of maintaining invariants upon themselves. In this context, type safety means that it is impossible in normal operation to try and deserialize a document at a wrong type or address (note that all primitive query functions are polymorphic), and risk getting bogus data without errors being reported.

Marks a field available for sorting and range queries.

Indexing requires a ToKey instance. Apart from the provided Bool, Int and DateTime instances, there is an overlappable fallback instance based on converting the Show string representation to an IxKey by taking the first 4 or 8 bytes. This is good enough for primitive string sorting.

Unique fields act as primary keys, and can be queried with unique and updateUnique. The Hashable instance is used to generate the key.

For fields that need to be both unique and sortable, use Unique (Sortable a) rather then the other way around.

This class is used by the generic scrapper to extract indexable keys from the fields of a Document. There are instances for Reference, Sortable and Unique, a general Foldable instance, and a special one for Maybe that converts Nothing into 0, such that null values will be indexed too, and become queryable with filter.

Users are not expected to need writing instances for this class. They can rather be generated automatically with the DeriveAnyClass extension.

Class used by the generic key scrapper to extract indexing information from user types. Only an empty instance needs to be added in user code.

Data type used by the key scrapper to collect keys while traversing generically user types.

Used inside indexes and as arguments to query primitives.

Class used to convert indexable field & query argument values to keys.

It is possibly needed for users to write instances for IxKey-convertible primitive types, other then the provided Bool, Int and DateTime.

Primary key for a document.

Allocated by functions in the Database.Muesli.IdSupply module.

Key extracted from non-reference fields for building a SortIndex.

Key extracted from unique fields for building a UniqueIndex

Key generated by Property for building a SortIndex or a FilterIndex.

With the OverloadedStrings extension, or directly using the IsString instance, field names specified in query arguments are converted to Property. An IxKey is computed by hashing the property name together with the TypeRep of the phantom argument. This key is used in indexes.

A newtype wrapper around UTCTime that has Show, Serialize and ToKey instances.

Type for exceptions thrown by the database. During normal operation these should never be thrown.

Transaction ids are auto-incremented globally. See mkNewTransactionId.

Address in the abstract data file of a serialized document's data.

These addresses are allocated by functions in the module Database.Muesli.Allocator.

Size of a serialized document's data.