Schema to type

Hey guys, how do you make a schema item as a static type ? for example, this thing:

export const orderItems = pgTable("orderItem", {
        product_id: serial("product_id").notNull(),
        order_id: serial("order_id").notNull(),
        quantity: integer("quantity")

    },
    (orderItems) => ({
            compoundKey: primaryKey(orderItems.product_id, orderItems.order_id),
        }
    ))

export const orderItems = pgTable("orderItem", {
        product_id: serial("product_id").notNull(),
        order_id: serial("order_id").notNull(),
        quantity: integer("quantity")

    },
    (orderItems) => ({
            compoundKey: primaryKey(orderItems.product_id, orderItems.order_id),
        }
    ))

{
productId: number,
order_id: number,
quantity: number
}

{
productId: number,
order_id: number,
quantity: number
}

14 Replies

Angelelz•15mo ago

There are several ways: https://orm.drizzle.team/docs/goodies#type-api

Goodies - DrizzleORM

Drizzle ORM | %s

wailrothOP•15mo ago

bless you

wailrothOP•15mo ago

so I made it, but have few questions about it: for example in this code

wailrothOP•15mo ago

Angelelz•15mo ago

You can't use the ifering helpers like that You have to use the table itself There are plenty of examples in the docs

wailrothOP•15mo ago

oh, okay lol i'll rtfm and stop annoying you

Angelelz•15mo ago

I can help, no problem, but so can the docs

wailrothOP•15mo ago

yeah; but I didnt see something in the doc: can you set custom messages if something is missing... Like I did in my custom schema?

Angelelz•15mo ago

I don't understand the question, can you write an example?

wailrothOP•15mo ago

and, like for example, with my useForm, values are string then I check the type with a transform

wailrothOP•15mo ago

but in this case, with and insertProductSchema, it doesnt work and like that:

productId: z.number({
        required_error: "Product id is required",
        invalid_type_error: "Product id is a number"
    }).optional(),
    name: z.string({
        required_error: "A name is required",
        invalid_type_error: "Name should be a string"
    }),

productId: z.number({
        required_error: "Product id is required",
        invalid_type_error: "Product id is a number"
    }).optional(),
    name: z.string({
        required_error: "A name is required",
        invalid_type_error: "Name should be a string"
    }),

Angelelz•15mo ago

If you're using zod, you might want to take a look at drizzle-zod It's an extension to create zod schemas for validations, seems like it's exactly what you're looking for

wailrothOP•15mo ago

In my case, I think i'll be forced to create a new schema because, for example, my product take in attributes "price" which is a number, but with the 'useForm' hook, you'r forced to take a string, then transform it to a number, then submit it like that for example:

    price: z.string({
        required_error: "A price is required",
        invalid_type_error: "Price should be a number"
    }).transform((val) => parseInt(val)),

    price: z.string({
        required_error: "A price is required",
        invalid_type_error: "Price should be a number"
    }).transform((val) => parseInt(val)),

Angelelz•15mo ago

That's up to you how you want to handle your business logic. You can detach your front end logic from whatever you have in your backend by writing your own schema validators But you could pull the schema from the db using drizzle-zod, end then transform it however you want for your application zod has helpers for that

Gaming

Programming

Schema to type