BitBadger.Documents/src/Postgres

BitBadger.Documents.Postgres

This package provides a lightweight document library backed by PostgreSQL. It also provides streamlined functions for traditional ADO.NET functionality where relational data is required. Both C# and F# have first-class implementations.

Features

• Select, insert, update, save (upsert), delete, count, and check existence of documents, and create tables and indexes for these documents
• Automatically generate IDs for documents (numeric IDs, GUIDs, or random strings)
• Address documents via ID, via comparison on any field, via equality on any property (using JSON containment, on a likely indexed field), or via condition on any property (using JSON Path queries)
• Access documents as your domain models (POCOs)
• Use Task-based async for all data access functions
• Use building blocks for more complex queries

Upgrading from v3

There is a breaking API change for ByField (C#) / byField (F#), along with a compatibility namespace that can mitigate the impact of these changes. See the migration guide for full details.

Getting Started

Once the package is installed, the library needs a data source. Construct an NpgsqlDataSource instance, and provide it to the library:

// C#
using BitBadger.Documents.Postgres;

//...
// Do not use "using" here; the library will handle disposing this instance
var data = new NpgsqlDataSourceBuilder("connection-string").Build();
Postgres.Configuration.UseDataSource(data);

// F#
open BitBadger.Documents.Postgres

// ...
// Do not use "use" here; the library will handle disposing this instance
let dataSource = // same as above ....

Configuration.useDataSource dataSource
// ...

By default, the library uses a System.Text.Json-based serializer configured to use the FSharp.SystemTextJson converter. To provide a different serializer (different options, more converters, etc.), construct it to implement IDocumentSerializer and provide it via Configuration.useSerializer. If custom serialization makes the serialized Id field not be Id, that will also need to be configured.

Using

Retrieve all customers:

// C#; parameter is table name
// Find.All type signature is Func<string, Task<List<TDoc>>>
var customers = await Find.All<Customer>("customer");

// F#
// Find.all type signature is string -> Task<'TDoc list>
let! customers = Find.all<Customer> "customer"

Select a customer by ID:

// C#; parameters are table name and ID
// Find.ById type signature is Func<string, TKey, Task<TDoc?>>
var customer = await Find.ById<string, Customer>("customer", "123");

// F#
// Find.byId type signature is string -> 'TKey -> Task<'TDoc option>
let! customer = Find.byId<string, Customer> "customer" "123"

(keys are treated as strings or numbers depending on their defintion; however, they are indexed as strings)

Count customers in Atlanta (using JSON containment):

// C#; parameters are table name and object for containment query
// Count.ByContains type signature is Func<string, TCriteria, int>
var customerCount = await Count.ByContains("customer", new { City = "Atlanta" });

// F#
// Count.byContains type signature is string -> 'TCriteria -> Task<int>
let! customerCount = Count.byContains "customer" {| City = "Atlanta" |}

Delete customers in Chicago: (no offense, Second City; just an example...)

// C#; parameters are table name and JSON Path expression
// Delete.ByJsonPath type signature is Func<string, string, Task>
await Delete.ByJsonPath("customer", "$.City ? (@ == \"Chicago\")");

// F#
// Delete.byJsonPath type signature is string -> string -> Task<unit>
do! Delete.byJsonPath "customer" """$.City ? (@ == "Chicago")"""

More Information

The project site has full details on how to use this library.