为什么Redis中没有有序的hashmap? [英] Why there is no ordered hashmap in Redis?

问题描述

Redis 数据类型 包括 排序集 和其他用于键值存储的必要数据结构.但我想知道为什么它没有像 Java 的 TreeMap 或 C++ 的 std::map 那样的排序映射.我认为底层数据结构将与排序集大致相似，因为两者都应该是平衡的二叉搜索树.

必须有一些用例，我们必须根据键以特定顺序存储键值对.但是当前的 sorted set 仅用于根据 score 存储 key 的目的.

解决方案

必须有一些用例，我们必须根据键以特定顺序存储键值对

由于Redis键是二进制字符串，我假设您提到的具体顺序是字典顺序(具体来说，键与memcmp函数进行比较).在这种情况下，您可以使用 SORTED SET 轻松实现 C++ 的 std::map.您可以通过 2 个步骤实现此目的:

<块引用>

使用 Redis 的排序集构建 std::set

如果 SORTED SET 中的 2 个元素具有相同的分数，则它们按字典顺序排列.因此，为了构建一个 std::set，只需给 SORTED SET 中的所有成员分配相同的分数:

zadd std::set 0 czadd std::set 0 azadd std::set 0 b//因为所有这些成员都有相同的分数，//结果按字典顺序排列://a b czrange std::set 0 -1//以下命令将失败，因为 'c' 已经存在.zadd std::set 0 c

从 Redis 2.8 开始，它支持一些命令来操作词典范围，以便您可以构建类似于 std::set::lower_bound 或 std::set::upper_bound

的内容

//类似于lower_bound:查找不小于b的所有成员zrangebylex std::set [b +//类似于upper_bound:查找所有大于b的成员zrangebylex std::set (b +

<块引用>

用一个值映射集合中的每个键

既然你已经得到了一个std::set，那么把key映射到一个值，你就可以得到一个std::map.

设置一个值_a设置 b value_b设置 c value_c

<块引用>

将这两个步骤结合在一起

您可以将整个工作包装到一个 lua 脚本中以具有 内置 std::map.并像这样使用它:

redis-cli --eval map.lua map_name ，键值

Redis Data types includes sorted set and other necessary data-structures for key-value storage. But I wonder why it doesn't have any sorted map like Java's TreeMap or C++'s std::map. I think the underlying data-structure would be mostly similar of sorted set as both are supposed to be balanced binary search tree.

There must be some use-cases where we have to store key-value pair in specific order according to key. But current sorted set only serves the purpose of storing key according to score.

解决方案

There must be some use-cases where we have to store key-value pair in specific order according to key

Since Redis keys are binary strings, I assume that the specific order you mentioned, is lexicographical order (specifically, keys are compared with the memcmp function). In that case, you can easily implement a C++'s std::map with SORTED SET. You can achieve this with 2 steps:

Build a std::set with Redis' Sorted Set

If 2 elements in a SORTED SET have the same score, they're ordered in lexicographical order. So in order to build a std::set, just give all members in a SORTED SET with the same score:

zadd std::set 0 c
zadd std::set 0 a
zadd std::set 0 b

// since all these members have the same score,
// the result is lexicographical ordered:
// a b c
zrange std::set 0 -1

// the following command will fail, since 'c' already exists.
zadd std::set 0 c

Since Redis 2.8, it supports some commands to operate on the lexicographical ranges, so that you can build something similar to std::set::lower_bound, or std::set::upper_bound

// something similar to lower_bound: find all members not less than b
zrangebylex std::set [b +
// something similar to upper_bound: find all members greater than b
zrangebylex std::set (b +

Map each key in the set with a value

Since you already get a std::set, then map the key with a value, you can get a std::map.

set a value_a
set b value_b
set c value_c

Combine these 2 steps together

You can wrap the whole work into a lua script to have a built-in std::map. And use it like this:

redis-cli --eval map.lua map_name , key value

这篇关于为什么Redis中没有有序的hashmap?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！