Indexing semantic versions in RocksDB
blog.aawadia.dev
Indexing semantic versions in RocksDB
1–10 of 18 posts
Re: Indexing semantic versions in RocksDB
#2I don't know anything about rocksdb, but this approach on surface level seems like it could be very slow? Wouldn't it be more efficient to encode the semver in a format more suitable to sorting
Re: Indexing semantic versions in RocksDB
#3> "The technical primitive data structure here is a hashmap where the keys are sorted."
..somewhere, this person's algorithms teacher is pondering about his life choices.
Re: Indexing semantic versions in RocksDB
#4> The technical primitive data structure here is a hashmap where the keys are sorted.
Why not a tree based map instead?
Re: Indexing semantic versions in RocksDB
#5I am curious, I recently wrote a naive hashmap for C. I am curious about iterating in insert and sort order.
Is it possible for a hash function to maintain a sort relationship to it's input and output?
Re: Indexing semantic versions in RocksDB
#6I am curious, I recently wrote a naive hashmap for C. I am curious about iterating in insert and sort order. Is it possible for a hash function to maintain a sort relationship to it's input and output?
If you iterate a python dictionary it will return the keys in insert order.
For sort order you need to sort separately.
Re: Indexing semantic versions in RocksDB
#7I don't know anything about rocksdb, but this approach on surface level seems like it could be very slow? Wouldn't it be more efficient to encode the semver in a format more suitable to sorting
Sure, the old performance/productivity tradeoff.
This quickly solves the problem then you can iterate for performance.
Re: Indexing semantic versions in RocksDB
#8> The technical primitive data structure here is a hashmap where the keys are sorted. Why not a tree based map instead?
HashMap and sorted keys sounds like a contradiction. I think they just use "HashMap" as a generic term for "Map" and it's actually a tree based map.
Re: Indexing semantic versions in RocksDB
#9Using RocksDB here seems fairly insane unless you need to keep (at least) several billion constantly-updating versions sorted. Otherwise you can just use SortedMap and regular Java comparators.
Re: Indexing semantic versions in RocksDB
#10> The technical primitive data structure here is a hashmap where the keys are sorted. Why not a tree based map instead?
HashMap and sorted keys sounds like a contradiction. I think they just use "HashMap" as a generic term for "Map" and it's actually a tree based map.
It is an LSMT, but I cannot imagine why they would need one here.