PostgreSQL · Delete
Delete on a preloaded schema: keys only, Strict
Delete on a schema preloaded at service setup, in three scenarios: the convenient setup (keys only) deleting by the key the preload found, with the table it added because it is referenced; Strict refusing a delete matched by a column that wasn't preloaded; and the same delete on a service that isn't Strict.
Set up for this page
Besides the sample itself, it uses SampleContext.cs (the dbContext every sample is given); SampleOutput.cs (Step and Show, which print); SampleContext.Application.cs (dbContext.Service, the application's ARX service); SampleCatalog.cs (the tables' ids, keys, lookups and display names), in the files below.
Before scenario 1 the sample runs this, so the service and the rows the scenarios use are there:
foreach (var (first, email) in new[] { ("Dora", "dora.delete@example.com"), ("Dev", "dev.delete@example.com"), ("Dee", "dee.delete@example.com") })
{
var contact = new ArxEntity("contact");
contact["first_name"] = first;
contact["last_name"] = "Deleteme";
contact["email"] = email;
await service.CreateAsync(contact);
}1 KeysOnly: delete by the key the preload found
Keys only: contact.email is unique, so it is a key; account comes too (contact.account_id references it).
Already in scope: connectionString (string)
var tables = new ArxPreLoadTableRegister()
.PreLoadTable("contact");
var schema = await new ArxPreLoadTable(tables, ArxPreLoadMode.KeysOnly).LoadAsync(connectionString);
var service = new ArxPgDbService(connectionString, schema);
foreach (var table in schema.Tables)
{
Console.WriteLine($"{table.Name,-8} {table.Reason}; keys {string.Join(" ", table.Keys)}");
}
var dora = new ArxEntity("contact");
dora["email"] = "dora.delete@example.com";
Console.WriteLine($"deleted {await service.DeleteAsync(dora, new ArxDeleteOptions { Match = ArxMatch.Key("email") })} row");Check it in the database
Run it in any SQL tool connected to the database, after the code has run.
SELECT count(*)
FROM contact
WHERE email = 'dora.delete@example.com';Expect: 0.
2 Strict: a delete matched by a column that wasn't preloaded
Strict with only email preloaded: deleting every contact named "Deleteme" is refused before the database.
Already in scope: connectionString (string)
var tables = new ArxPreLoadTableRegister()
.PreLoadTable("contact", ["email"]);
var schema = await new ArxPreLoadTable(tables).LoadAsync(connectionString);
var strict = new ArxPgDbService(connectionString, schema) { Strict = true };
var byLastName = new ArxEntity("contact");
byLastName["last_name"] = "Deleteme";
try
{
await strict.DeleteAsync(byLastName, new ArxDeleteOptions { Match = ArxMatch.Columns("last_name") });
}
catch (ArxValidationException exception) when (exception.Code == ArxErrorCode.NotPreloaded)
{
Console.WriteLine(exception.Message);
}
var byEmail = new ArxEntity("contact");
byEmail["email"] = "dev.delete@example.com";
Console.WriteLine($"by the preloaded email: deleted {await strict.DeleteAsync(byEmail, new ArxDeleteOptions { Match = ArxMatch.Key("email") })} row");Check it in the database
Run it in any SQL tool connected to the database, after the code has run.
SELECT email
FROM contact
WHERE last_name = 'Deleteme';Expect: 1 row, dee.delete@example.com: the delete by last_name was refused, dev's by email ran.
3 Not Strict: the same delete works
The same schema, not Strict: last_name isn't preloaded, and the delete by it still works.
Already in scope: connectionString (string)
var tables = new ArxPreLoadTableRegister()
.PreLoadTable("contact", ["email"]);
var schema = await new ArxPreLoadTable(tables).LoadAsync(connectionString);
var service = new ArxPgDbService(connectionString, schema);
var byLastName = new ArxEntity("contact");
byLastName["last_name"] = "Deleteme";
Console.WriteLine($"deleted {await service.DeleteAsync(byLastName, new ArxDeleteOptions { Match = ArxMatch.Columns("last_name") })} row");Check it in the database
Run it in any SQL tool connected to the database, after the code has run.
SELECT count(*)
FROM contact
WHERE last_name = 'Deleteme';Expect: 0.
Good to know
Under Strict, every column that finds the rows to delete must be preloaded, so an endpoint can only delete by the columns you chose. Create Advanced6 shows the preload itself and its JSON.
Run the complete sample
The code above comes from one program, built in a fresh project and run against the database. To run it yourself, here is every file and step.
The program, its files and how to run it
Run it locally
- .NET SDK: .NET SDK 10.0.100 or later 10.0.x (target
net10.0). - Create the project:
Terminal dotnet new console -n ArxSample -f net10.0 cd ArxSample - Install the packages:
Terminal dotnet add package ARX.SDK.PostgreSql --version 1.0.0-rc.1 - Database: a PostgreSQL 16 (tested with 16.15) database with the sample schema.
- Configuration: set
ARX_SAMPLES_POSTGRESQLto its connection string. Nothing is hard-coded. - Source code: replace
Program.csand add the 5 other files below, with the file names shown. - Run it:
Terminal dotnet run - Expected output: below.
The files
Runs the sample. Generated for this page; it reads the connection string from the environment.
using ArxSamples.PostgreSql;
using ArxSamples.PostgreSql.Operations.Delete.AdvancedCases;
// The connection string comes from the environment, never from the code (see "Configure the connection").
var connectionString = Environment.GetEnvironmentVariable("ARX_SAMPLES_POSTGRESQL")
?? throw new InvalidOperationException("Set the ARX_SAMPLES_POSTGRESQL environment variable to your PostgreSQL connection string.");
await new Advanced3_DynamicTableSchemaServiceSetup().RunAsync(new SampleContext(connectionString));
The sample, as it is in the repository.
using ARX.SDK;
using ARX.SDK.PostgreSql.Services;
using ARX.SDK.Services;
using static ArxSamples.PostgreSql.SampleOutput;
namespace ArxSamples.PostgreSql.Operations.Delete.AdvancedCases;
/// <summary>
/// Delete on a schema preloaded at service setup, in three scenarios: the convenient setup (keys only)
/// deleting by the key the preload found, with the table it added because it is referenced; Strict refusing a delete
/// matched by a column that wasn't preloaded; and the same delete on a service that isn't Strict.
/// </summary>
/// <remarks>
/// Under Strict, every column that finds the rows to delete must be preloaded, so an endpoint can only delete by the
/// columns you chose. Create Advanced6 shows the preload itself and its JSON.
/// </remarks>
public sealed class Advanced3_DynamicTableSchemaServiceSetup : ISample
{
public string Name => "delete-advanced-3-dynamic-table-schema-service-setup";
public string Title => "Delete on a preloaded schema: keys only, Strict";
public async Task RunAsync(SampleContext dbContext)
{
foreach (var (first, email) in new[] { ("Dora", "dora.delete@example.com"), ("Dev", "dev.delete@example.com"), ("Dee", "dee.delete@example.com") })
{
var contact = new ArxEntity("contact");
contact["first_name"] = first;
contact["last_name"] = "Deleteme";
contact["email"] = email;
await dbContext.Service.CreateAsync(contact);
}
await Scenario1_DeleteByTheKeyThePreloadFoundAsync(dbContext);
await Scenario2_StrictRefusesAColumnNotPreloadedAsync(dbContext);
await Scenario3_NotStrictTheSameDeleteWorksAsync(dbContext);
}
/// <summary>1. Keys only: contact.email is unique, so it is a key; account comes too (contact.account_id references it).</summary>
private static async Task Scenario1_DeleteByTheKeyThePreloadFoundAsync(SampleContext dbContext)
{
Step("1. KeysOnly: delete by the key the preload found");
var tables = new ArxPreLoadTableRegister()
.PreLoadTable("contact");
var schema = await new ArxPreLoadTable(tables, ArxPreLoadMode.KeysOnly).LoadAsync(dbContext.ConnectionString);
var service = new ArxPgDbService(dbContext.ConnectionString, schema);
foreach (var table in schema.Tables)
{
Show($"{table.Name,-8} {table.Reason}; keys {string.Join(" ", table.Keys)}");
}
var dora = new ArxEntity("contact");
dora["email"] = "dora.delete@example.com";
Show($"deleted {await service.DeleteAsync(dora, new ArxDeleteOptions { Match = ArxMatch.Key("email") })} row");
/* ===========================================================
* [DIRECT QUERY CHECK]
* -----------------------------------------------------------
* SELECT count(*)
* FROM contact
* WHERE email = 'dora.delete@example.com';
* -- Expect: 0.
* ===========================================================
*/
}
/// <summary>2. Strict with only email preloaded: deleting every contact named "Deleteme" is refused before the database.</summary>
private static async Task Scenario2_StrictRefusesAColumnNotPreloadedAsync(SampleContext dbContext)
{
Step("2. Strict: a delete matched by a column that wasn't preloaded");
var tables = new ArxPreLoadTableRegister()
.PreLoadTable("contact", ["email"]);
var schema = await new ArxPreLoadTable(tables).LoadAsync(dbContext.ConnectionString);
var strict = new ArxPgDbService(dbContext.ConnectionString, schema) { Strict = true };
var byLastName = new ArxEntity("contact");
byLastName["last_name"] = "Deleteme";
try
{
await strict.DeleteAsync(byLastName, new ArxDeleteOptions { Match = ArxMatch.Columns("last_name") });
}
catch (ArxValidationException exception) when (exception.Code == ArxErrorCode.NotPreloaded)
{
ShowError(exception);
}
var byEmail = new ArxEntity("contact");
byEmail["email"] = "dev.delete@example.com";
Show($"by the preloaded email: deleted {await strict.DeleteAsync(byEmail, new ArxDeleteOptions { Match = ArxMatch.Key("email") })} row");
/* ===========================================================
* [DIRECT QUERY CHECK]
* -----------------------------------------------------------
* SELECT email
* FROM contact
* WHERE last_name = 'Deleteme';
* -- Expect: 1 row, dee.delete@example.com: the delete by last_name was refused, dev's by email ran.
* ===========================================================
*/
}
/// <summary>3. The same schema, not Strict: last_name isn't preloaded, and the delete by it still works.</summary>
private static async Task Scenario3_NotStrictTheSameDeleteWorksAsync(SampleContext dbContext)
{
Step("3. Not Strict: the same delete works");
var tables = new ArxPreLoadTableRegister()
.PreLoadTable("contact", ["email"]);
var schema = await new ArxPreLoadTable(tables).LoadAsync(dbContext.ConnectionString);
var service = new ArxPgDbService(dbContext.ConnectionString, schema);
var byLastName = new ArxEntity("contact");
byLastName["last_name"] = "Deleteme";
Show($"deleted {await service.DeleteAsync(byLastName, new ArxDeleteOptions { Match = ArxMatch.Columns("last_name") })} row");
/* ===========================================================
* [DIRECT QUERY CHECK]
* -----------------------------------------------------------
* SELECT count(*)
* FROM contact
* WHERE last_name = 'Deleteme';
* -- Expect: 0.
* ===========================================================
*/
}
}
What every sample is given: the connection string (and the ISample interface the samples implement).
namespace ArxSamples.PostgreSql;
/// <summary>One documentation sample: a scenario a real application has, shown end to end.</summary>
public interface ISample
{
/// <summary>The name to run it alone: <c>dotnet run -- <name></c>.</summary>
string Name { get; }
/// <summary>One line on what it shows.</summary>
string Title { get; }
Task RunAsync(SampleContext dbContext);
}
/// <summary>
/// What a sample is given. Every sample gets the connection string; the application's service and catalog
/// (SampleContext.Application.cs) and the plain Npgsql read-back helper (SampleContext.Direct.cs) live in their own
/// files, so a sample needs only the files it uses.
/// </summary>
public sealed partial class SampleContext
{
public SampleContext(string connectionString)
{
ConnectionString = connectionString;
}
public string ConnectionString { get; }
}
How the samples print: Step for each scenario, Show for what it did, ShowError for an error caught on purpose.
using ARX.SDK;
namespace ArxSamples.PostgreSql;
/// <summary>How the samples print: a numbered step for each scenario, then what it did.</summary>
public static class SampleOutput
{
/// <summary>The start of a scenario: <c>Step("1. By the id Create returned")</c>.</summary>
public static void Step(string text) => Console.WriteLine($" > {text}");
/// <summary>What the scenario did.</summary>
public static void Show(string text) => Console.WriteLine($" {text}");
/// <summary>An ARX error the scenario caught on purpose.</summary>
public static void ShowError(ArxException exception)
=> Console.WriteLine($" caught {exception.GetType().Name}: {exception.Message}");
}
The application's ARX service and its schema catalog, created once and shared by the samples.
using ARX.SDK.PostgreSql.Services;
using ARX.SDK.Services;
namespace ArxSamples.PostgreSql;
/// <summary>The application's ARX service and its schema catalog, shared by the real and advanced samples.</summary>
public sealed partial class SampleContext
{
private ArxSchemaCatalog? _catalog;
private ArxPgDbService? _service;
/// <summary>The application's schema catalog: table ids, keys and selection sets (SampleCatalog.cs).</summary>
public ArxSchemaCatalog Catalog => _catalog ??= SampleCatalog.Build();
/// <summary>
/// One service for the whole application. It holds no connection: every call opens one from the Npgsql pool and
/// closes it, unless you pass your own connection or transaction.
/// </summary>
public ArxPgDbService Service => _service ??= new ArxPgDbService(ConnectionString, Catalog);
}
The application's schema catalog: every table's id column and keys, and its selection sets.
using ARX.SDK.Services;
using System.Text.Json;
using System.Text.Json.Serialization;
namespace ArxSamples.PostgreSql;
/// <summary>An account's industry. Stored in account.industry (smallint) as the enum's number.</summary>
public enum Industry : short
{
Retail = 1,
Manufacturing = 2,
Healthcare = 3,
Finance = 4,
Technology = 5
}
/// <summary>
/// The application's schema catalog: every table's id column and keys, and every picklist (selection set),
/// defined once.
/// </summary>
/// <remarks>
/// <para>
/// <b>Ids.</b> Every table the application writes is declared with its one id column: its name, its .NET type, and
/// whether the database generates it (<c>ArxIdColumn.Generated</c>: identity, sequence, default) or the caller supplies
/// it (<c>ArxIdColumn.Supplied</c>: a natural id, a client-generated Guid). A table with no id column, such as a log or a
/// table identified by several columns together, is declared <c>ArxIdColumn.NoIdColumn</c>. ARX never reads ids from
/// the database: the developer knows the schema, and an undeclared table is refused (ARX1016).
/// </para>
/// <para>
/// <b>Keys.</b> Optional business keys: one or more columns whose values identify a record, such as an account number,
/// or a region and a SKU together. Create doesn't use them (the database's unique constraints do the enforcing); Update
/// and Delete will be able to find rows by them. <c>ValidateSchemaAsync</c> checks ids and keys against the database
/// when you ask it to.
/// </para>
/// <para>
/// <b>Selection sets.</b> /// A selection set (picklist, option set) is a column that only accepts a fixed list of codes, each with a display
/// text: an account's status (ACTIVE, ON_HOLD, CLOSED), a contact's preferred channel (E, P, S). The catalog tells ARX,
/// for each set:
/// </para>
/// <list type="bullet">
/// <item><b>Name</b>: what an <see cref="ArxSelectionSetValue"/> refers to, for example <c>new ArxSelectionSetValue("AccountStatus", "ON_HOLD")</c>.</item>
/// <item><b>Kind</b>: what the codes are, matching the column: Integer, String, Character, Boolean, Decimal, Guid, or a
/// PostgreSQL enum type (DatabaseEnum).</item>
/// <item><b>Multiplicity</b>: Single (one code), or several codes stored as an array (Array), bit flags (Flags), a
/// delimited string (Delimited) or a JSON array (Json).</item>
/// <item><b>Options</b>: each code with its display text and order. A retired option (<c>IsActive: false</c>) stays
/// readable in old rows, but can't be written any more.</item>
/// </list>
/// <para>
/// Build the catalog once at startup and give it to the service: <c>new ArxPgDbService(connectionString, SampleCatalog.Build())</c>.
/// ARX checks every <see cref="ArxSelectionSetValue"/> against it in memory, before any SQL is sent, and stores the code
/// itself in the column: no lookup table, no extra query. A plain value (a string or number that isn't an
/// ArxSelectionSetValue) isn't checked.
/// </para>
/// <para>
/// The catalog is plain code, so the options can come from wherever you keep them: written out here, taken from a .NET
/// enum (<c>AddEnum</c>), loaded from JSON (<see cref="SetsJson"/> and <see cref="FromJson"/> below), or loaded at
/// startup from your own option table and passed to <c>new ArxSelectionSetDefinition(name, kind, options)</c>. ARX
/// never reads option tables or files itself.
/// </para>
/// </remarks>
public static class SampleCatalog
{
/// <summary>
/// Selection sets kept as JSON, as an application might keep them in a file, a settings store or a database column.
/// Here the JSON is in the class itself; in an application, read it from wherever you keep it
/// (<c>File.ReadAllText("selection-sets.json")</c>, configuration, an HTTP call) and pass it to <see cref="FromJson"/>.
/// "multiplicity" and "active" are optional (Single and true).
/// </summary>
public const string SetsJson = """
[
{
"name": "LeadSource",
"kind": "String",
"options": [
{ "value": "WEB", "text": "Website", "order": 0 },
{ "value": "REFERRAL", "text": "Referral", "order": 1 },
{ "value": "EVENT", "text": "Trade show", "order": 2 },
{ "value": "COLD_CALL", "text": "Cold call", "order": 3, "active": false }
]
},
{
"name": "ContactRating",
"kind": "Integer",
"multiplicity": "Single",
"options": [
{ "value": 1, "text": "Poor" },
{ "value": 2, "text": "Fair" },
{ "value": 3, "text": "Good" },
{ "value": 4, "text": "Very good" },
{ "value": 5, "text": "Excellent" }
]
}
]
""";
public static ArxSchemaCatalog Build()
{
var builder = ArxSchemaCatalog.Create()
// ---------- tables: id column and keys ----------
// Generated bigint ids (bigserial): the database fills them in. The keys are the unique columns that also
// identify a record.
.Table("account", ArxIdColumn.Generated<long>("id")).Key("account", "account_number")
.Table("contact", ArxIdColumn.Generated<long>("id")).Key("contact", "email")
// Retrieve: a contact's account_id reads back as a reference to its account, named by the account's
// name; a contact is named by its first and last names.
.Lookup("contact", "account_id", "account")
.DisplayName("account", "name")
.DisplayName("contact", new[] { "first_name", "last_name" }, " ")
.Table("product_category", ArxIdColumn.Generated<long>("id")).Key("product_category", "name")
.Table("product", ArxIdColumn.Generated<long>("id")).Key("product", "sku")
.Table("sales_order", ArxIdColumn.Generated<long>("id")).Key("sales_order", "order_number")
.Table("order_line", ArxIdColumn.Generated<long>("id"))
.Table("data_type_showcase", ArxIdColumn.Generated<long>("id"))
// A natural id: the caller gives the ISO country code.
.Table("country", ArxIdColumn.Supplied<string>("code"))
// A generated uuid id (DEFAULT gen_random_uuid()).
.Table("device", ArxIdColumn.Generated<Guid>("id")).Key("device", "serial_number")
// No id column: a price is identified by its region and SKU together, a compound key.
.Table("price_list", ArxIdColumn.NoIdColumn).Key("price_list", "region", "sku")
// No primary key at all: the same action can be logged twice.
.Table("activity_log", ArxIdColumn.NoIdColumn)
// A partitioned table without an id: a visit is a fact, never updated (Retrieve Advanced8).
.Table("web_visit", ArxIdColumn.NoIdColumn)
// ---------- selection sets ----------
// From a .NET enum: the set is named after the enum ("Industry"), the codes are its numbers (Retail = 1, …),
// the display texts are its member names. Column: account.industry smallint.
.AddEnum<Industry>()
// Written out: String codes with display texts, single choice. Column: account.status text.
.Add("AccountStatus", ArxSelectionSetKind.String, ("ACTIVE", "Active"), ("ON_HOLD", "On hold"), ("CLOSED", "Closed"))
// Multiple choice, stored as an array of codes. RESELLER is retired: rows that have it still read back,
// but new writes can't use it. Column: account.segments text[].
.Add(new ArxSelectionSetDefinition("CustomerSegment", ArxSelectionSetKind.String, new[]
{
new ArxSelectionSetOption("SMB", "Small business", 0),
new ArxSelectionSetOption("ENTERPRISE", "Enterprise", 1),
new ArxSelectionSetOption("PUBLIC", "Public sector", 2),
new ArxSelectionSetOption("RESELLER", "Reseller", 3, IsActive: false)
}, ArxSelectionSetMultiplicity.Array))
// Single-character codes. Column: contact.preferred_channel char(1).
.Add("ContactChannel", ArxSelectionSetKind.Character, ('E', "Email"), ('P', "Phone"), ('S', "Text message"))
// The labels of a PostgreSQL enum type: the value is written with a cast to that type.
// Column: data_type_showcase.mood showcase_mood.
.Add(new ArxSelectionSetDefinition("Mood", ArxSelectionSetKind.DatabaseEnum, new[]
{
new ArxSelectionSetOption("happy", "Happy", 0),
new ArxSelectionSetOption("neutral", "Neutral", 1),
new ArxSelectionSetOption("sad", "Sad", 2)
}, databaseTypeName: "showcase_mood"));
// Loaded from JSON: LeadSource (contact.lead_source text) and ContactRating (contact.rating smallint).
foreach (var set in FromJson(SetsJson))
{
builder.Add(set);
}
return builder.Build();
}
/// <summary>Reads selection set definitions from JSON (the shape of <see cref="SetsJson"/>).</summary>
public static IEnumerable<ArxSelectionSetDefinition> FromJson(string json)
{
var options = new JsonSerializerOptions
{
PropertyNameCaseInsensitive = true,
Converters = { new JsonStringEnumConverter() } // "String", "Integer", "Array", … as names
};
var sets = JsonSerializer.Deserialize<List<JsonSet>>(json, options)
?? throw new InvalidOperationException("The selection set JSON is empty.");
foreach (var set in sets)
{
yield return new ArxSelectionSetDefinition(
set.Name,
set.Kind,
set.Options.Select(x => new ArxSelectionSetOption(Code(x.Value, set.Kind), x.Text, x.Order, x.Active)),
set.Multiplicity,
set.DatabaseTypeName);
}
}
/// <summary>A JSON value as a code of the set's kind: JSON has no char, Guid or 64-bit integer of its own.</summary>
private static object Code(JsonElement value, ArxSelectionSetKind kind) => kind switch
{
ArxSelectionSetKind.Integer => value.GetInt64(),
ArxSelectionSetKind.Decimal => value.GetDecimal(),
ArxSelectionSetKind.Boolean => value.GetBoolean(),
ArxSelectionSetKind.Guid => value.GetGuid(),
ArxSelectionSetKind.Character => value.GetString()![0],
_ => value.GetString()! // String and DatabaseEnum
};
private sealed record JsonSet(
string Name,
ArxSelectionSetKind Kind,
List<JsonOption> Options,
ArxSelectionSetMultiplicity Multiplicity = ArxSelectionSetMultiplicity.Single,
string? DatabaseTypeName = null);
private sealed record JsonOption(JsonElement Value, string Text, int Order = 0, bool Active = true);
}
Expected output
What dotnet run printed when the sample was verified, on a freshly created sample schema. Ids differ when the tables already hold rows.
> 1. KeysOnly: delete by the key the preload found
contact requested; keys (email)
account added: referenced by contact.account_id; keys (account_number)
deleted 1 row
> 2. Strict: a delete matched by a column that wasn't preloaded
caught ArxValidationException: ARX1018 NotPreloaded: Column 'last_name' of 'contact' wasn't preloaded, and the service is Strict. Preloaded: id, email. Add it to the ArxPreLoadTable, or turn Strict off.
by the preloaded email: deleted 1 row
> 3. Not Strict: the same delete works
deleted 1 row