support dynamic conditions, In the form of struct or map.

[Neightly]

Bun's goal is to help you write idiomatic SQL, not to hide it behind awkward constructs.

It is good for static code, when table schemas are known at compile time.

What about dynamic tables created at runtime?
Unknown about table name and field name, how I can use Where family methods to construct conditions?

Field-based query conditions usually exist as a whole, so support for structs would be nice.

[j2gg0s]

Can you provide more specific use cases?

[Neightly]

// dummy fields known at runtime
var Main1118 = func() *dynamic.DomainRepository {
	Main := dynamic.NewDomainRepository("test", "jux_main_1118", "main1118")
	Main.AddUint64("SystemId", dynamic.WithRaw(`json:"system_id" bun:"system_id"`))
	Main.AddString("TypeName", dynamic.WithRaw(`json:"typename" bun:"type_name"`))
	Main.AddInt32("RowIndex", dynamic.WithRaw(`json:"rowindex" bun:"row_index"`))
	Main.AddInt8("Status", dynamic.WithRaw(`json:"status" bun:"status"`))
	return Main
}()

// create table
_ = Main1118.CreateTable(db)

// cond and ptr both point the same value, cond for reflect, ptr for static
cond, ptr = Main1118.StructValue(map[string]any{
    // conditions are different each time.
    "SystemId": uint64(99), // "test"."jux_main_1118"."system_id" = 99
    "TypeName": "XXXX", // "test"."jux_main_1118"."type_name" = 'XXXX'
})
_, records, err := Main1118.Slice(db, cond, ptr)

[j2gg0s]

Thank you for the additional use cases.

But I am a supporter of static strong typing. Let’s see what other maintainers think.