$rand 本身不能生成字符串型优惠码，必须配合 $concat、$arrayElemAt、$toString 等操作符组合使用；单独用 $rand 只能得到 0–1 的浮点数，需乘以字符池长度取整后查表拼接才能生成如 "XK7M9P" 的6位字母数字优惠码。
直接说结论：$rand 本身不能生成字符串型优惠码，必须配合 $concat、$arrayElemAt、$toString 等操作符组合使用；单独用 $rand 只能得到 0–1 的浮点数。
为什么不能直接用 $rand 生成优惠码
$rand 是纯数值生成器，返回值形如 0.37284912345678901，没有字符映射、无长度控制、不可逆向构造字母数字组合。想得到类似 "XK7M9P" 这样的优惠码，它只是整个链条里的“随机种子”环节。
常见错误现象包括：
- 误写
{$concat: [{$rand: {}}]}→ 报错：Expression $rand is not a string - 试图用
$rand直接替换$sample做随机文档选取 → 完全走不通，语义不匹配 - 在
$project中漏掉类型转换，导致后续$concat失败
用 $rand + $project 构建 6 位字母数字优惠码
核心思路：把 $rand 当作均匀分布的随机源，乘以字符池长度取整，再用 $arrayElemAt 查表拼接。
实操建议如下：
- 先定义一个字符池数组，比如
["A","B","C",...,"9"]（共 36 个字符），放在$let或硬编码进表达式 - 对每一位，执行
{$floor: {$multiply: [{$rand: {}}, 36]}}得到 0–35 的索引 - 用
{$arrayElemAt: [字符池数组, 索引]}取出对应字符 - 6 次重复上述逻辑，最后用
$concat合并
简化版示例（6 位）：
db.coupons.aggregate([
{
$project: {
code: {
$concat: [
{ $arrayElemAt: [["0","1","2","3","4","5","6","7","8","9","A","B","C","D","E","F","G","H","I","J","K","L","M","N","O","P","Q","R","S","T","U","V","W","X","Y","Z"], { $floor: { $multiply: [{$rand: {}}, 36] } }] },
{ $arrayElemAt: [["0","1","2","3","4","5","6","7","8","9","A","B","C","D","E","F","G","H","I","J","K","L","M","N","O","P","Q","R","S","T","U","V","W","X","Y","Z"], { $floor: { $multiply: [{$rand: {}}, 36] } }] },
{ $arrayElemAt: [["0","1","2","3","4","5","6","7","8","9","A","B","C","D","E","F","G","H","I","J","K","L","M","N","O","P","Q","R","S","T","U","V","W","X","Y","Z"], { $floor: { $multiply: [{$rand: {}}, 36] } }] },
{ $arrayElemAt: [["0","1","2","3","4","5","6","7","8","9","A","B","C","D","E","F","G","H","I","J","K","L","M","N","O","P","Q","R","S","T","U","V","W","X","Y","Z"], { $floor: { $multiply: [{$rand: {}}, 36] } }] },
{ $arrayElemAt: [["0","1","2","3","4","5","6","7","8","9","A","B","C","D","E","F","G","H","I","J","K","L","M","N","O","P","Q","R","S","T","U","V","W","X","Y","Z"], { $floor: { $multiply: [{$rand: {}}, 36] } }] },
{ $arrayElemAt: [["0","1","2","3","4","5","6","7","8","9","A","B","C","D","E","F","G","H","I","J","K","L","M","N","O","P","Q","R","S","T","U","V","W","X","Y","Z"], { $floor: { $multiply: [{$rand: {}}, 36] } }] }
]
}
}
}
])
注意：每次调用 {$rand: {}} 都是独立随机事件，不会复用同一值 —— 这正是我们需要的。
$sample 更适合“从已有优惠码池中随机选”，而非“生成新码”
如果你已有 10 万条预生成的唯一优惠码存于 codes 集合，且只需每次取一条未使用的，$sample 是更高效、更可靠的选择：
-
$sample在满足条件（首阶段、集合 >100 条、采样数 $rand 计算 - 避免了并发写入时因
$rand重复导致的优惠码冲突问题 - 可配合
$match: {used: false}和$set: {used: true}实现原子性领取
典型组合：
db.codes.aggregate([
{ $match: { used: false } },
{ $sample: { size: 1 } },
{ $set: { used: true } },
{ $merge: { into: "codes", on: "_id", whenMatched: "replace" } }
])
这种模式下，$rand 完全不需要出场。
真正容易被忽略的一点：用 $rand 批量生成大量优惠码时，MongoDB 不保证全局唯一性 —— 它不校验重复，也不去重。你得在外层加唯一索引（code: 1 + unique: true），并准备好捕获 DuplicateKey 错误做重试。这不是聚合管道能解决的问题。
