如何在没有两个查询的情况下在 Knex 中执行此操作答案

【问题标题】：How can I do this in Knex without two queries如何在没有两个查询的情况下在 Knex 中执行此操作
【发布时间】：2020-08-13 02:30:44
【问题描述】：

我想在不使用两个查询的情况下进行此搜索：

app.db('users').where({
            id: req.userData.id
        }).first().then(result => {

            if (!result) {
                throw new CompleteError('Usuário não encontrado.', 404)
            }

            app.db('addresses').where({
                id: result.address_id
            }).first().then(address_result => {
                if (address_result) {
                    result.address = address_result
                    res.json(result)
                } else {
                    throw new CompleteError('Falha ao buscar endereço do usuário.', 500)
                }
            }).catch(e => {
                return next(e)
            })

        }).catch(e => {
            return next(e)
        })

如果其中一个搜索失败或没有找到任何寄存器，搜索进度将停止并抛出错误。数据库是postgresql。如果此方法没有任何替代方法，请帮助如何改进此代码，我是初学者，我需要您的选择。请，对不起我的英语。

结果一定是这样的：

{
  "id": 21,
  "name": "Sérgio",
  "cpf": "11111111111",
  "email": "sergio11@gmail.com",
  "pass": "$2a$10$Jx7wAd3QSMFlkXi3g0eVOOoW8o65OgDZVQ7wTWuxquytVhUhrUf/.",
  "phone": "5544313113131",
  "last_access": "2020-08-11T10:14:20.207Z",
  "created_at": "2020-08-11T10:14:20.207Z",
  "deleted_at": null,
  "address_id": 21,
  "address": {
    "id": 21,
    "street": "Street Manu",
    "number": "851",
    "complement": "A",
    "city": "Mari Hill",
    "state": "Para",
    "country": "Brazil",
    "last_access": "2020-08-11T10:14:19.837Z",
    "created_at": "2020-08-11T10:14:19.837Z",
    "deleted_at": null
  }
}

【问题讨论】：

标签： node.js postgresql knex.js

【解决方案1】：

您可以使用 JOIN 执行此查询，然后重新排列结果。

您要查找的查询如下所示：

SELECT *
FROM users
    LEFT JOIN addresses ON addresses.id=users.address_id;

（我对 MySQL/MariaDB 比较熟悉，所以语法可能有点不同）。

LEFT JOIN 表示如果找到用户但没有找到地址，则地址列将为空。

这可以翻译成下面的 Knex 表达式：

app.db('users')
    .leftJoin('addresses', 'addresses.id', '=', 'users.address_id')
    .select(
        'users.id',
        'users.name',
        'users.cpf',
        'users.email',
        'users.pass',
        'users.phone',
        'users.last_access',
        'users.created_at',
        'users.deleted_at',
        'users.address_id',
        'addresses.id AS addr_id',
        'addresses.street AS addr_street',
        'addresses.number AS addr_number',
        'addresses.complement AS addr_complement',
        'addresses.city AS addr_city',
        'addresses.state AS addr_state',
        'addresses.country AS addr_country',
        'addresses.last_access AS addr_last_access',
        'addresses.created_at AS addr_created_at',
        'addresses.deleted_at AS addr_deleted_at'
    )
    .first()
    .then((result) => {
        // no result at all? no user found
        if (!result) {
            throw new CompleteError('Usuário não encontrado.', 404);
        }

        // no addr_id? user found but not the address
        if (!result.addr_id) {
            throw new CompleteError('Falha ao buscar endereço do usuário.', 500);
        }

        // result contains all the fields from BOTH tables as a flat object
        // you'll need to rearrange things
        return {
            // assign properties to new names...
            id: result.id,
            // and so on...
        };
    });

请注意，我已在 select() 调用中重命名了 addresses 中的所有列。这是因为 Knex 不会在带有 JOIN 子句的查询中分配新的列名，除非您明确命名列，否则您将得到重复的列名（并且会造成很多混乱）。

【讨论】：