Jest.mock()失效：测试仍调用@databricks/sql原模块求助

核心问题分析 #

你的测试超时且mock未生效，主要有两个原因：

• 模拟的@databricks/sql API结构与原模块不匹配：原模块的方法（如connect、openSession）直接返回Promise，但你的mock返回的是带有promise方法的对象，导致await无法正确解析，最终超时。
• 模块级实例化时机：待测试文件中databricksClient = new DBSQLClient()是在模块加载时执行的，若mock未提前生效，会直接实例化原模块的类。

具体修复步骤 #

1. 修正mock的API结构，匹配原模块的Promise返回 #

将mock中所有方法改为直接返回Promise，而非嵌套promise方法的对象。修改后的@databricks/sql mock如下：

jest.mock("@databricks/sql", () => {
  return {
    DBSQLClient: function () {
      return {
        // connect直接返回resolved Promise
        connect: () => Promise.resolve(),
        openSession: () => Promise.resolve({
          executeStatement: () => Promise.resolve({
            fetchAll: () => Promise.resolve(mockResponse),
            close: () => Promise.resolve()
          })
        }),
        close: () => Promise.resolve()
      };
    },
  };
});

这样await databricksClient.connect()就能正确解析，不会卡住。

2. 调整测试文件的模块加载顺序 #

确保jest.mock在加载待测试模块前执行（Jest会自动提升jest.mock到顶部，但显式放在最顶部更稳妥）：

// 先定义mock
jest.mock("@databricks/sql", () => { /* ... */ });
jest.mock("aws-sdk", () => { /* ... */ });

// 再导入依赖和待测试模块
import mockResponse from "./mockResponse.json";
const { handler } = require("../../../backend/function/dataLakeConnect/src/index");

3. 验证模块导出结构（可选） #

如果mock仍未生效，可在待测试文件中临时添加打印，确认@databricks/sql的导出结构：

console.log(require("@databricks/sql"));

若原模块的DBSQLClient是默认导出，需调整mock为：

jest.mock("@databricks/sql", () => {
  return {
    default: function () { /* ... */ }
  };
});

同时待测试文件的导入也要改为const DBSQLClient = require("@databricks/sql").default;。

4. 清理模块缓存（可选） #

如果之前的测试残留了模块缓存，可在测试前强制清除：

beforeEach(() => {
  jest.resetModules();
});

修复后的完整测试文件示例 #

jest.mock("@databricks/sql", () => {
  return {
    DBSQLClient: function () {
      return {
        connect: () => Promise.resolve(),
        openSession: () => Promise.resolve({
          executeStatement: () => Promise.resolve({
            fetchAll: () => Promise.resolve(mockResponse),
            close: () => Promise.resolve()
          })
        }),
        close: () => Promise.resolve()
      };
    },
  };
});

jest.mock("aws-sdk", () => {
  return {
    SecretsManager: function () {
      return {
        getSecretValue: ({ SecretId }) => {
          if (SecretId === "credentials/databricks/read-write") {
            return {
              promise: () => Promise.resolve({
                ARN: "x",
                Name: "credentials/databricks/read-write",
                VersionId: "x",
                SecretString:
                  '{"DATABRICKS_TOKEN":"test","DATABRICKS_SERVER_HOSTNAME":"test.test.com","DATABRICKS_HTTP_PATH":"/pathy/pathpath/mcgee"}',
                VersionStages: ["x"],
                CreatedDate: "x",
              })
            };
          } else {
            throw new Error("mock error");
          }
        },
      };
    },
  };
});

import mockResponse from "./mockResponse.json";
const { handler } = require("../../../backend/function/dataLakeConnect/src/index");

const testEvent = {
  arguments: {
    page: 1,
    clientId: "test",
    startTime: "2022-09-26 00:00:00",
    endTime: "2022-09-27 00:00:00",
  },
};

describe("dataLakeConnect", () => {
  test("should return expected array", async () => {
    const res = await handler(testEvent);
    expect(res).toEqual(mockResponse);
  });
});

内容的提问来源于stack exchange，提问作者Alisha Edington