I was processing compressed .jsonl files recently (JSON lines format). I found that lzma gave a much better compression than gzip or bzip2, which helps for archival costs, but it's challenging to work with as software support is lacking. I do duckdb processing which supports gzip transparently. There's an extension for bzip2, but not for lzma or bzip3.
I ended up using gzip because it's best supported by the software I use and most likely to have support in software I adopt. But it gave the worst compression results of the options I tried. These bzip3 numbers certainly give me FOMO...
For that type of structured data (logs and such), a custom dictionary can be extremely effective. Zstd among others support generating a custom dictionary. You just run zstd --train over the data first, and then feed that in when you run zstd. For example: I found ~10 gigabytes of Usenet headers compress to ~700 MB using Zstd and 1 MB shared dictionary -- and that's with each header individually compressed, so o(1) lookup time.
Back in my data hoarder days, I downloaded one of those torrents that had all the world's books in it. It was dunno how many terabytes, but way more than I had HDDs.
So I stripped out formatting, got rid of dupes, and tried out zstd, which was the hot new thing, along with the dictionary feature you describe, figuring it'd help. It didn't. I tried having one per book, one per multiple books, one for the whole archive.
It didn't work, or the gains were so marginal that I ended up scrapping the approach.
So it's not impossible that it can work, but stuff like regular json already compresses extremely well, I haven't found a scenario where it's a major boon.
When I studied at school, I used ZFS with lz4 enabled on my working machine. During that times I had a task of parsing Wikipedia's data. I had enough brain cells to find compressed dumps and download them with aria2 but not enough to leave the file compressed. I ran a decompressor. It'd been taking longer than I expected so I went out to walk a dog.
Imagine how fast me and the dog ran back 30 minutes later when I realized how cooked I was. I only had 10 GB left on my disks after I downloaded that 20 GB file. This decompressed file would have blown the machine up. I was terrified to find a frozen system with no storage space left.
Instead, the process finished and `df -h` reported 8 GB of the free space left. Files were decompressed. I could `less` them! That made no sense! Only many many minutes later I finally figured out to run a `zfs get compressratio` command which showed ZFS successfully and transparently recompressed everything on the fly. That was too impressive for that teenager and he never switched to a different file system.
Solaris has had so many cool features, like ZFS or doors. What I liked about ZFS is you coul make snapshot, which is basically the solution to how to treat data files a single, cheap to access unit, yet still use standard apis for file management, great for containerizing apps, making copies for experiments, or shipping stuff. Node.js just received this as a bespoke, app-level feature. But these things are too many to count, and make a ton of sense if you know how filesystems actually work. useful Also copy-on-write, temp overlays.
Pretrained dictionaries have never been intended to help with book sized or bigger compression. zstd automatically learns the most efficient dictionary it can within a few kilobytes. Pretrained dictionaries are only useful when you're independently compressing very small records.
Note that the dictionary options are only needed to improve compression ratios when compressing lots of small messages. If you have a bigger file (eg a tar file of Usenet messages) the regular Zstd compression will build a good dictionary without additional options.
The thing zstd got really right is fast decompression. For write-once read-never data like backups lzma (aka xz/7zip/lzip) is great. But it takes forever to decompress. On zstd I can get good compression while decompressing the file only marginally slower than reading the uncompressed file from SSD
Writing your files directly into a compressed stream and decompressing on the fly has become almost a standard workflow for any files I'm going to read and write sequentially anyways. No need for the data to ever exist uncompressed on the file system. Previous formats never did that for me because they either had too much overhead or too little gain, often both
But zstd is super tunable. Where gzip gives you compression levels from 1 to 9, zstd gives you up to 22 for ultra compression and negative compression levels for ultra fast. The ultra fast options so fast that they are great as a substitute for memcpy if your CPU is already waiting for other things, like DRAM.
No it’s not. The pace of improvement of CPU compute speed is far greater than that of DRAM throughput. And in fact compression algorithms geared towards speed aims to outperform memcpy (on suitable machines).
zstd has a built-in benchmark mode to compare different compression levels, e.g. `zstd -b1 -e9 [FILE]` to test levels 1 to 9 (try up to 22 if you have enough spare time)
Duckdb supports loading and saving to zstd for all it's base loading/saving formats csv/tsv/json/jsonlines, but, for good or bad, those are solid compression.
Under most r/w workloads, using parquet/lance/vortex/native-duckdb, with their built-in columnar compression will result in more performance AND space savings. Non-solid compression. Then, the query engine can push down your query predicate to a column row group level, instead of forcing it to decompress the entire dataset to operate.
Practical example: duckdb has syntax - https://duckdb.org/docs/lts/data/multiple_files/overview - to glob multiple files at once, but that really only works if you're applying push down query predicates instead of re-decompressing your entire data set per SELECT. I would say for most dataset, even 20%+ size is worth not having to decompress (or even download!) the entire dataset, to figure out if something fits the predicate.
After all, if you have to download and decompress the dataset back again to operate, then the "space savings" are gone.
It isn't really the go-to compression format, because it isn't ubiquitous like gzip and zip, there are a variety of compression tools out there for different purposes, and there is image/audio/video compression. There is also specialized compression like what git does with its rolling hashes. I think of it as there not being a go-to compression format.
A go-to thing means it's a sensible default choice and has no little to no downsides (versus not using compression), it doesn't mean it's the best for everything.
Until now the go-to has been DEFLATE (gzip and zip) but zstd is definitely competing against it because it is better in almost every way.
I ended up using gzip because it's best supported by the software I use and most likely to have support in software I adopt. But it gave the worst compression results of the options I tried. These bzip3 numbers certainly give me FOMO...