v2.8.0

SqlHydra.Query v2.8.0

☀️.NET 9 Support (thanks, @EverybodyKurts!)
☀️Added support for SQL Server output in insert and update builders which allows you to return one or more inserted columns.

open SqlHydra.Query
open SqlHydra.Query.SqlServerExtensions 

let insertPerson (row: Person) =
  task {
    let! createDate, updateDate =
        insertTask openContext {
            for p in dbo.Person do
            entity row
            output (e.CreateDate, e.UpdateDate)
        }

    return updateDate
  }

open SqlHydra.Query
open SqlHydra.Query.SqlServerExtensions 

let updateRaceway (req: SaveRequest) = 
  updateAsync openContext {
      for rw in dbo.RACEWAYS do
      set rw.DRAWING_AREA req.DrawingArea
      set rw.UPDATED_DATE DateTime.Now
      where (rw.JOB_ID = jobId && req.RUN_NUMBER = runNo)
      output rw.UPDATED_DATE
  }

SqlHydra.Cli v2.8.1

☀️.NET 9 Support (thanks, @EverybodyKurts!)
☀️Fixed so that SQL Server HierarchyId column (added in v2.7.0) is not added to the generated primitive types function unless HierarchyId is actually used in the generated schema (since it requires an extra NuGet package).

v2.7.0

SqlHydra.Cli v2.7.0

😺 SQL Server generation now supports HierarchyId column type. (Thanks to @michelbieleveld for contributing this feature!)

SqlHydra.Query v2.7.0

😺 Add new areEqual and notEqual query functions to support querying against the HierarchyId column type.
It is necessary to use these functions since the HierarchyId type overloads the = operator to return a SqlBoolean instead of a bool, which breaks the query syntax.

Usage example:

    let node = child.GetAncestor(1)

    let! result = 
        selectTask ctx { 
            for row in ext.HierarchyIdSupport do
            where (row.Id = id_parent && areEqual row.Hierarchy node)
            select row.Hierarchy
            toList
        }

v2.6.0

SqlHydra.Cli v2.6.0

😺 Generating types for a netstandard project will now automatically generate "legacy mode" DateTime and TimeSpan instead of using DateOnly and TimeOnly types.

😺 New selectTask and selectAsync computation expressions are now generated to work with your generated HydraReader.Read and table types.

Previously, you had to manually pass in the generated HydraReader.Read method to your selectTask and selectAsync builders:

open SqlHydra.Query
open AdventureWorks.Generated

let getErrorNumbers () =
    selectTask HydraReader.Read openContext {
        for e in dbo.ErrorLog do
        select e.ErrorNumber
    }

Now, you can use the generated versions which have the HydraReader.Read method baked-in:

open SqlHydra.Query
open AdventureWorks.Generated 
open AdventureWorks.Generated.HydraBuilders // Generated builders are here

let getErrorNumbers () =
    selectTask openContext {
        for e in dbo.ErrorLog do
        select e.ErrorNumber
    }

v2.5.3

SqlHydra.Cli

• Added timestamp mapping for SQL Server #103
• Merged MySql code gen #101

v2.6.0-beta.1 - MySQL Beta Support

This beta release adds type generation support to SqlHydra.Cli for MySQL. (Thanks to @RJSonnenberg for the contribution!)

v2.5.0

SqlHydra.Query v2.5.0

• In a select query, specifying a table or tables using the select keyword now explicitly selects all the columns in the table record; previously, it would issue a select tbl.* query. It is now recommended to always explicitly select a table. (If you do not explicitly select, it will generate a select tbl.* style query which will be less performant.)

Ex:

❌ This will issue a SELECT * query which will result in a table scan which will result in slightly worse performance.

    let! results = 
        selectTask' openContext {
            for p in Person.Person do
            take 10
        }

✅ This will explicity select all columns in the Person table, which will result in slightly better performance.

    let! results = 
        selectTask' openContext {
            for p in Person.Person do
            take 10
            select p
        }

SqlHydra.Cli v2.5.0

• The generated HydraReader now filters out any columns that do not exist in the selected table record(s). (This fixes a bug that could happen when a column was added to a table and the types were not regenerated.

NOTE: You must upgrade both SqlHydra.Query and SqlHydra.Cli at the same time to v2.5.x.

v2.4.1

SqlHydra.Query

• Added implicit conversions to ContextType.

The ContextType (Shared, Create, CreateAsync and CreateTask) is now implicitly converted.
This allows you to either pass in a QueryContext or a function that returns a QueryContext to your selectTask and selectAsync expressions. This is a huge quality-of-life improvement, IMO!

See Select Builders in readme for more examples.

Old:

let getErrorNumbers () =
    selectAsync HydraReader.Read (Create openContext) {
        for e in dbo.ErrorLog do
        select e.ErrorNumber
    }

Becomes:

let getErrorNumbers () =
    selectAsync HydraReader.Read openContext {
        for e in dbo.ErrorLog do
        select e.ErrorNumber
    }

Old:

let getErrorNumbers (ctx: QueryContext) =
    selectAsync HydraReader.Read (Shared ctx) {
        for e in dbo.ErrorLog do
        select e.ErrorNumber
    }

Becomes:

let getErrorNumbers (ctx: QueryContext) =
    selectAsync HydraReader.Read ctx {
        for e in dbo.ErrorLog do
        select e.ErrorNumber
    }

If you create a shortcut select CE that embeds the generated HydraReader.Read function, it becomes even more succinct:

let selectTask' ct = selectTask HydraReader.Read ct

let distinctCustomerNames (ctx: QueryContext) = 
    selectTask' ctx {
        for c in SalesLT.Customer do
        select (c.FirstName, c.LastName)
        distinct
    }

v2.4.0

SqlHydra.Cli

• Rewrote code generation using Fabulous.AST!

• Added new mutable_properties option to toml which makes all record properties mutable. 🙀 (defaults to false.)

• Added new nullable_property_type option to toml which allows nullable columns to be generated as either F# option types (default) or System.Nullable properties. The latter can be useful for C# interop or CRUD scenarios where you want to data bind generated types directly to UI controls. (Generally, I would recommend mapping generated types to domain entities or DTOs, but it's nice to have the option to do things in a quick and dirty way.)

  • mutable_properties = true
  • nullable_property_type = "Nullable"

• Improved CLI Output

  • The entire TOML configuration is now printed to CLI
  • TOM include/exclude filters are now printed to CLI, along with the number of tables generated (post-filter) and the total number of tables found (pre-filter) to the console. #88
    

• All providers will now ignore any tables / views with no columns (just in case)

• Postgres "time with time zone" columns now generate DateTimeOffset properties

• Postgres now generates materialized views (requires net8 and npgsql v8 or greater). #84

• Updated CLI to latest "Microsoft.Data.SqlClient" to fix warnings

SqlHydra.Query

• Changes to support the new System.Nullable column/properties in the Linq query syntax

• Changes to support the new System.Nullable column/properties query parameters

• Adds CreateTask and CreateAsync cases to the ContextType discriminated union. #89

• Exposed KataQuery property on all queries (the resulting type of query builders) to make it possible to manipulate the underlying SqlKata query for inserts, updates and selects before executing the query.

• Added head custom operation to the SelectBuilder for queries where you know there should always be at least one result. (Before you had to use tryHead, and then manually extract the value.)

v2.4.0-beta.6

SqlHydra.Query

• Adds CreateTask and CreateAsync cases to the ContextType discriminated union.

SqlHydra.Cli

• Updates to the latest version of Fabulous.AST for code generation

v2.4.0-beta.5

SqlHydra.Cli

• Now prints the active filters, number of tables generated (post-filter) and the total number of tables found (pre-filter) to the console. #88

SqlHydra.Query

• No changes.