documentation/docs/data/data-provider/index.md
import SupportedDataProviders from "@site/src/partials/data-provider/supported-data-providers.md";
Data provider acts as a data layer for your app, making HTTP requests and encapsulating how the data is retrieved. The methods of these requests are then consumed by Refine via data hooks.
You don’t need to worry about creating data providers from scratch, as Refine offers built-in data provider support for the most popular API providers.
:::simple Good to know
:::
<Image src="https://refine.ams3.cdn.digitaloceanspaces.com/website/static/img/guides-and-concepts/providers/data-provider/api-consuming-flow.webp" />To activate the data provider in Refine, we have to pass the dataProvider to the <Refine /> component.
import { Refine } from "@refinedev/core";
import dataProvider from "./dataProvider";
const App = () => (
<Refine
/* ... */
dataProvider={dataProvider}
/>
);
Refer to the Data Provider tutorial for more information and usage examples →
Refine allows you to use multiple data providers in your app. All you need to do is pass key and value pairs to the dataProvider prop of the <Refine /> component. In the pair object, the key corresponds to the data provider name, and the value corresponds to the data provider itself.
When defining multiple data providers, default key is required for defining the default data provider.
Here is an example which uses multiple data providers:
setInitialRoutes(["/posts"]);
setRefineProps({ Sider: () => null });
// visible-block-start
import { Refine, useList } from "@refinedev/core";
import routerProvider from "@refinedev/react-router";
import dataProvider from "@refinedev/simple-rest";
import { BrowserRouter, Route, Routes } from "react-router";
import { Collapse, Tag } from "antd";
const API_URL = "https://api.fake-rest.refine.dev";
const FINE_FOODS_API_URL = "https://api.finefoods.refine.dev";
interface IPost {
id: number;
title: string;
status: "published" | "draft" | "rejected";
}
interface IProduct {
id: number;
name: string;
price: number;
}
const App = () => {
return (
<BrowserRouter>
<Refine
routerProvider={routerProvider}
// highlight-start
dataProvider={{
// `default` is required to determine the default data provider
default: dataProvider(API_URL),
fineFoods: dataProvider(FINE_FOODS_API_URL),
}}
// highlight-end
resources={[
{
// highlight-next-line
// Refine will use the `default` data provider for this resource
name: "posts",
list: "/posts",
},
{
name: "products",
meta: {
// highlight-start
// Refine will use the `fineFoods` data provider for this resource
dataProviderName: "fineFoods",
// highlight-end
},
},
]}
>
<Routes>
<Route path="/posts" element={<PostList />} />
</Routes>
</Refine>
</BrowserRouter>
);
};
const PostList = () => {
const {
result: { data: posts },
} = useList<IPost>({
resource: "posts",
// highlight-start
// Data provider can be selected through props
dataProviderName: "default",
// highlight-end
});
// highlight-start
// We've defined the data provider for this resource as "fineFoods" in its config so we don't need to pass it here
const {
result: { data: products },
} = useList<IProduct>({ resource: "products" });
// highlight-end
console.log({
posts,
products,
});
return (
<Collapse defaultActiveKey={["products"]}>
<Collapse.Panel header="Posts" key="posts">
{posts?.map((post) => (
<div
key={post.title}
style={{
display: "flex",
flexDirection: "row",
gap: "0.5rem",
marginBottom: "0.25rem",
}}
>
{post.title}
<Tag>{post.status}</Tag>
</div>
))}
</Collapse.Panel>
<Collapse.Panel header="Products" key="products">
{products?.map((product) => (
<div
key={product.name}
style={{
display: "flex",
flexDirection: "row",
gap: "0.5rem",
marginBottom: "0.25rem",
}}
>
{product.name}
<Tag>{product.price / 10}</Tag>
</div>
))}
</Collapse.Panel>
</Collapse>
);
};
// visible-block-end
render(<App />);
You can pick data providers in two ways:
You can either use the dataProviderName prop in data hooks and data-related components/functions:
useTable({
meta: {
dataProviderName: "example",
},
});
Or use the meta.dataProviderName property in your resource config, which will be the default data provider but can be overridden in data hooks and components:
const App = () => (
<Refine
dataProvider={{
default: defaultDataProvider,
example: exampleDataProvider,
}}
resources={[
{
// `default` data provider will be used for this resource
name: "posts",
},
{
name: "products",
// `exampleDataProvider` data provider will be used for this resource
meta: { dataProviderName: "exampleDataProvider" },
},
]}
/>
);
Data provider's methods are expected to return a promise, meaning that they are async and can be used to create a data provider. Refine consumes these data provider methods using data hooks, which are used for CRUD actions like creating a new record, listing a resource or deleting a record, etc.
import { DataProvider } from "@refinedev/core";
const dataProvider: DataProvider = {
// required methods
getList: ({ resource, pagination, sorters, filters, meta }) => Promise,
create: ({ resource, variables, meta }) => Promise,
update: ({ resource, id, variables, meta }) => Promise,
deleteOne: ({ resource, id, variables, meta }) => Promise,
getOne: ({ resource, id, meta }) => Promise,
getApiUrl: () => "",
// optional methods
getMany: ({ resource, ids, meta }) => Promise,
createMany: ({ resource, variables, meta }) => Promise,
deleteMany: ({ resource, ids, variables, meta }) => Promise,
updateMany: ({ resource, ids, variables, meta }) => Promise,
custom: ({ url, method, filters, sorters, payload, query, headers, meta }) =>
Promise,
};
Refer to the Data Provider tutorial for more information and usage examples→
The getList method is used to get a list of resources with sorting, filtering, and pagination features. It takes resource, sorters, pagination, and, filters as parameters. And it returns both data and total fields, regardless of the data provider used.
Refine will consume this method using the useList or useInfiniteList data hook.
x-total-count header is commonly used to get the row count.response.data.pageInfo.total.This documentation provides only illustrative examples. It's up to the data provider to determine how to best source the total count.
The method signature remains the same, and Refine expects a consistent format:
getList: async ({ resource, pagination, sorters, filters, meta }) => {
const { currentPage, pageSize } = pagination ?? {};
// Adjust request parameters to meet the requirements of your API
const response = await apiClient.get(`/${resource}`, {
params: { _page: currentPage, _limit: pageSize },
});
// The total row count could be sourced differently based on the provider
const total = response.headers["x-total-count"] ?? response.data.length;
return {
data: response.data,
total,
};
};
:::tip
getList can also support cursor-based pagination. Refer to related section in the useInfiniteList documentation for more information.
:::
Parameter Types:
| Name | Type |
|---|---|
| resource | string |
| pagination? | Pagination |
| sorters? | CrudSorting |
| filters? | CrudFilters |
| meta? | MetaDataQuery |
The create method creates a new record with the resource and variables parameters.
Refine will consume this method using the useCreate data hook.
create: async ({ resource, variables, meta }) => {
// You can handle the request according to your API requirements.
return {
data,
};
};
Parameter Types
| Name | Type | Default |
|---|---|---|
| resource | string | |
| variables | TVariables | {} |
| meta? | MetaDataQuery |
TVariablesis a user defined type which can be passed touseCreateto typevariables.
The update method updates the record with the resource, id, and, variables parameters.
Refine will consume this method using the useUpdate data hook.
update: async ({ resource, id, variables, meta }) => {
// You can handle the request according to your API requirements.
return {
data,
};
};
Parameter Types:
| Name | Type | Default |
|---|---|---|
| resource | string | |
| id | BaseKey | |
| variables | TVariables | {} |
| meta? | MetaDataQuery |
TVariablesis a user defined type which can be passed touseUpdateto typevariables.
The deleteOne method delete the record with the resource and id parameters.
Refine will consume this method using the useDelete data hook.
deleteOne: async ({ resource, id, variables, meta }) => {
// You can handle the request according to your API requirements.
return {
data,
};
};
Parameter Types:
| Name | Type | Default |
|---|---|---|
| resource | string | |
| id | BaseKey | |
| variables | TVariables[] | {} |
| meta? | MetaDataQuery |
TVariablesis a user defined type which can be passed touseDeleteto typevariables.
The getOne method gets the record with the resource and id parameters.
Refine will consume this method using the useOne data hook.
getOne: async ({ resource, id, meta }) => {
// You can handle the request according to your API requirements.
return {
data,
};
};
Parameter Types:
| Name | Type | Default |
|---|---|---|
| resource | string | |
| id | BaseKey | |
| meta? | MetaDataQuery |
The getApiUrl method returns the apiUrl value.
Refine will consume this method using the useApiUrl data hook.
import { DataProvider } from "@refinedev/core";
export const dataProvider = (apiUrl: string): DataProvider => ({
getApiUrl: () => apiUrl,
// ...
});
An optional method named custom can be added to handle requests with custom parameters like URL, CRUD methods and configurations.
It's useful if you have non-standard REST API endpoints or want to make a connection with external resources.
Refine will consume this method using the useCustom data hook.
custom: async ({
url,
method,
filters,
sorters,
payload,
query,
headers,
meta,
}) => {
// You can handle the request according to your API requirements.
return {
data,
};
};
Parameter Types
| Name | Type |
|---|---|
| url | string |
| method | get, delete, head, options, post, put, patch |
| sorters? | CrudSorting |
| filters? | CrudFilters |
| payload? | {} |
| query? | {} |
| headers? | {} |
| meta? | MetaDataQuery |
Bulk actions are actions that can be performed on multiple items at once to improve speed and efficiency. They are commonly used in admin panels. They can be used for data import and export, and are also atomic, meaning that they are treated as a single unit.
If your API supports bulk actions, you can implement them in your data provider.
The getMany method gets the records with the resource and ids parameters. This method is optional, and Refine will use the getOne method to handle multiple requests if you don't implement it.
Refine will consume this getMany method using the useMany data hook.
getMany: async ({ resource, ids, meta }) => {
// You can handle the request according to your API requirements.
return {
data,
};
};
Parameter Types:
| Name | Type | Default |
|---|---|---|
| resource | string | |
| ids | [BaseKey] | |
| meta? | MetaDataQuery |
This method allows us to create multiple items in a resource. This method is optional, and Refine will use the create method to handle multiple requests if you don't implement it.
Refine will consume this createMany method using the useCreateMany data hook.
createMany: async ({ resource, variables, meta }) => {
// You can handle the request according to your API requirements.
return {
data,
};
};
Parameter Types:
| Name | Type | Default |
|---|---|---|
| resource | string | |
| variables | TVariables[] | {} |
| meta? | MetaDataQuery |
TVariablesis a user defined type which can be passed touseCreateManyto typevariables.
This method allows us to delete multiple items in a resource. This method is optional, and Refine will use the deleteOne method to handle multiple requests if you don't implement it.
Refine will consume this deleteMany method using the useDeleteMany data hook.
deleteMany: async ({ resource, ids, variables, meta }) => {
// You can handle the request according to your API requirements.
return {
data,
};
};
| Name | Type | Default |
|---|---|---|
| resource | string | |
| ids | [BaseKey] | |
| variables | TVariables[] | {} |
| meta? | MetaDataQuery |
TVariablesis a user defined type which can be passed touseDeleteManyto typevariables.
This method allows us to update multiple items in a resource. This method is optional, and Refine will use the update method to handle multiple requests if you don't implement it.
Refine will consume this updateMany method using the useUpdateMany data hook.
updateMany: async ({ resource, ids, variables, meta }) => {
// You can handle the request according to your API requirements.
return {
data,
};
};
| Name | Type | Default |
|---|---|---|
| resource | string | |
| ids | [BaseKey] | |
| variables | TVariables[] | {} |
| meta? | MetaDataQuery |
TVariablesis a user defined type which can be passed touseUpdateManyto typevariables.
Refine expects errors to be extended from HttpError.
Here is a basic example of how to implement error handling in your data provider.
import { DataProvider, HttpError } from "@refinedev/core";
export const dataProvider = (apiUrl: string): DataProvider => ({
getOne: async ({ resource, id }) => {
try {
const response = await fetch(`https://api.example.com/${resource}/${id}`);
// highlight-start
if (!response.ok) {
const error: HttpError = {
message: response.statusText,
statusCode: response.status,
};
return Promise.reject(error);
}
// highlight-end
return {
data: response.data,
};
} catch (error) {
// highlight-start
const error: HttpError = {
message: error?.message || "Something went wrong",
statusCode: error?.status || 500,
};
return Promise.reject(error);
// highlight-end
}
},
// ...
});
Also, the Axios interceptor can be used to transform the error from the response before Axios returns the response to your code.
// highlight-start
import axios from "axios";
import { DataProvider, HttpError } from "@refinedev/core";
// highlight-end
import { stringify } from "query-string";
// highlight-start
// Error handling with axios interceptors
const axiosInstance = axios.create();
axiosInstance.interceptors.response.use(
(response) => {
return response;
},
(error) => {
const customError: HttpError = {
...error,
message: error.response?.data?.message,
statusCode: error.response?.status,
};
return Promise.reject(customError);
},
);
// highlight-end
export const dataProvider = (apiUrl: string): DataProvider => ({
// Methods
});
When using APIs, you may wish to include custom parameters, such as a custom header. To accomplish this, you can utilize the meta field, which allows the sent parameter to be easily accessed by the data provider.
Here is an example of how to send a custom header parameter to the getOne method using meta:
We first need to send a custom header parameter to the getOne method using meta.
import { useOne } from "@refinedev/core";
useOne({
resource: "post",
id: "1",
meta: {
headers: {
"x-custom-header": "hello world",
},
},
});
Then we can get the meta parameter from the data provider:
import { DataProvider } from "@refinedev/core";
export const dataProvider = (apiUrl: string): DataProvider => ({
...
getOne: async ({ resource, id, variables, meta }) => {
// highlight-next-line
const { headers } = meta;
const url = `${apiUrl}/${resource}/${id}`;
// highlight-start
httpClient.defaults.headers = {
...headers,
};
// highlight-end
const { data } = await httpClient.get(url, variables);
return {
data,
};
},
});
Refine will consume:
getList method using the useList or useInfiniteList data hook.create method using the useCreate data hook.update method using the useUpdate data hook.deleteOne method using the useDeleteOne data hook.getOne method using the useOne data hook.getApiUrl method using the useApiUrl data hook.custom method using the useCustom data hook.getMany method using the useMany data hook.createMany method using the useCreateMany data hook.deleteMany method using the useDeleteMany data hook.updateMany method using the useUpdateMany data hook.Refer to the "Create Data Provider From Scratch" section in the tutorial for more information →
You can swizzle the data provider using the Refine CLI and customize it as needed.
In some cases, you may need to override the method of Refine data providers. The simplest way to do this is to use the Spread syntax
For example, Let's override the update function of the @refinedev/simple-rest. @refinedev/simple-rest uses the PATCH HTTP method for update, let's change it to PUT without forking the whole data provider.
import dataProvider from "@refinedev/simple-rest";
const simpleRestProvider = dataProvider("API_URL");
const myDataProvider = {
...simpleRestProvider,
update: async ({ resource, id, variables }) => {
const url = `${apiUrl}/${resource}/${id}`;
const { data } = await httpClient.put(url, variables);
return {
data,
};
},
};
<Refine dataProvider={myDataProvider}></Refine>;