That’s interesting! Can you say a little more? I find jq’s syntax and semantics to be simple and intuitive. It’s mostly dots, pipes, and brackets. It’s a lot like writing shell pipelines imo. And I tend to use it in the same way. Lots of one-time use invocations, so I spend more time writing jq filters than I spend reading them.
I suspect my use cases are less complex than yours. Or maybe jq just fits the way I think for some reason.
I dream of a world in which all CLI tools produce and consume JSON and we use jq to glue them together. Sounds like that would be a nightmare for you.
I'm not GP, I use jq all the time, but I each time I use it I feel like I'm still a beginner because I don't get where I want to go on the first several attempts. Great tool, but IMO it is more intuitive to JSON people that want a CLI tool than CLI people that want a JSON tool. In other words, I have my own preconceptions about how piping should work on the whole thing, not iterating, and it always trips me up.
Here's an example of my white whale, converting JSON arrays to TSV.
That whole map and from entries throws it off. It's not a good use for what you're doing. tsv expects a bunch of arrays, whereas you're getting a bunch of objects (with the header also being one) and then converting them to arrays. That is an unnecessary step and makes it a little harder to understand.
Thanks for sharing, this is much better, though I actually think it is the perfect example to explain something that is brain-slippery about jq
look at $cols | $cols
my brain says hmm that's a typo, clearly they meant ; instead of | because nothing is getting piped, we just have two separate statements. Surely the assignment "exhausts the pipeline" and we're only passing null downstream
the pipelining has some implicit contextual stuff going on that I have to arrive at by trial and error each time since it doesn't fit in my worldview while I'm doing other shell stuff
Honestly both of those make me do the confused-dog-head-tilt thing. I'd go for something sexp based, perhaps with infix composition, map, and flatmap operators as sugar.
I think the big problem is it's a tool you usually reach for so rarely you never quite get the opportunity to really learn it well, so it always remains in that valley of despair where you know you should use it, but it's never intuitive or easy to use.
It's not unique in that regard. 'sed' is Turing complete[1][2], but few people get farther than learning how to do a basic regex substitution.
I was just going to say, jq is like sed in that I only use 1% of it 99% of the time, but unlike sed in that I'm not aware of any clearly better if less ubiquitous alternatives to the 1% (e.g., Perl or ripgrep for simple regex substitutions in pipelines because better regex dialects).
Closest I've come, if you're willing to overlook its verbosity and (lack of) speed, is actually PowerShell, if only because it's a bit nicer than Python or JavaScript for interactive use.
CEL looks interesting and useful, though it isn't common nor familiar imo (not for me at least). Quoting from https://github.com/google/cel-spec
# Common Expression Language
The Common Expression Language (CEL) implements common
semantics for expression evaluation, enabling different
applications to more easily interoperate.
## Key Applications
- Security policy: organizations have complex infrastructure
and need common tooling to reason about the system as a whole
- Protocols: expressions are a useful data type and require
interoperability across programming languages and platforms.
Funny that everyone is linking the tools they wrote for themselves to deal with this problem. I am no exception. I wrote one that just lets you write JavaScript. Imagine my surprise that this extremely naive implementation was faster than jq, even on large files.
Like I did with regex some years earlier, I worked on a project for a few weeks that required constant interactions with jq, and through that I managed to lock in the general shape of queries so that my google hints became much faster.
Of course, this doesn't matter now, I just ask an LLM to make the query for me if it's so complex that I can't do it by hand within seconds.
To fix this I recently made myself a tiny tool I called jtree that recursively walks json, spitting out one line per leaf. Each line is the jq selector and leaf value separated by "=".
No more fiddling around trying to figure out the damn selector by trying to track the indentation level across a huge file. Also easy to pipe into fzf, then split on "=", trim, then pass to jq
LOL ... I can absolutely feel your pain. That's exactly why I created for myself a graphical approach. I shared the first version with friends and it turned into "ColumnLens" (ImGUI on Mac) app. Here is a use case from the healthcare industry: https://columnlens.com/industries/medical
Because the output you get can have hallucinations, which don’t happen with a deterministic tool. Furthermore, by getting the `jq` command you get something which is reusable, fast, offline, local, doesn’t send your data to a third-party, doesn’t waste a bunch of tokens, … Using an LLM to filter the data is worse in every metric.
I get that AI isn’t deterministic by definition, but IMHO it’s become the go-to response for a reason to not use AI, regardless of the use case.
I’ve never seen AI “hallucinate” on basic data transformation tasks. If you tell it to convert JSON to YAML, that’s what you’re going to get. Most LLMs are probably using something like jq to do the conversion in the background anyway.
AI experts say AI models don’t hallucinate, they confabulate.
Just because you haven't seen it hallucinate on these tasks doesn't mean it can't.
When I'm deciding what tool to use, my question is "does this need AI?", not "could AI solve this?" There's plenty of cases where its hard to write a deterministic script to do something, but if there is a deterministic option, why would you choose something that might give you the wrong answer? It's also more expensive.
The jq script or other script that an LLM generates is way easier to spot check than the output if you ask it to transform the data directly, and you can reuse it.
I already know that. That's why we have deterministic algorithms, to simplify that complexity.
You have much to learn, witty answers mean nothing here, particularly empty witty answers, which are no better than jokes. Maybe stand-up comedy is your call in life.
I appreciate performance as much as the next person; but I see this endless battle to measure things in ns/us/ms as performative.
Sure there are 0.000001% edge cases where that MIGHT be the next big bottleneck.
I see the same thing repeated in various front end tooling too. They all claim to be _much_ faster than their counterpart.
9/10 whatever tooling you are using now will be perfectly fine. Example; I use grep a lot in an ad hoc manner on really large files I switch to rg. But that is only in the handful of cases.
Whenever you have this kind of impressions on some development, here are my 2 cents: just think "I'm not the target audience". And that's fine.
The difference between 2ms and 0.2ms might sound unneeded, or even silly to you. But somebody, somewhere, is doing stream processing of TB-sized JSON objects, and they will care. These news are for them.
I remember when I was coming up on the command line and I'd browse the forums at unix.com. Someone would ask how to do a thing and CFAJohnson would come in with a far less readable solution that was more performative (probably replacing calls to external tools with Bash internals, but I didn't know enough then to speak intelligently about it now).
People would say, "Why use this when it's harder to read and only saves N ms?" He'd reply that you'd care about those ms when you had to read a database from 500 remote servers (I'm paraphrasing. He probably had a much better example.)
Turns out, he wrote a book that I later purchased. It appears to have been taken over by a different author, but the first release was all him and I bought it immediately when I recognized the name / unix.com handle. Though it was over my head when I first bought it, I later learned enough to love it. I hope he's on HN and knows that someone loved his posts / book.
Wow that takes me back. I used to lurk on unix.com when I was starting with bash and perl and would see CFAJohnson's terse one-liners all the time. I enjoyed trying my own approaches to compare performance, conciseness and readability - mainly for learning. Some of the awk stuff was quite illuminating in my understanding of how powerful awk could be. I remember trying different approaches to process large files at first with awk and then with Perl. Then we discovered Oracle's external tables which turned out to be clear winner. We have a lot more options now with fantastic performance.
Also as someone who looks at latency charts too much, what happens is a request does a lot in series and any little ms you can knock off adds up. You save 10ms by saving 10 x 1ms. And if you are a proxyish service then you are a 10ms in a chain that might be taking 200 or 300ms. It is like saving money, you have to like cut lots of small expenses to make an impact. (unless you move etc. but once you done that it is small numerous things thay add up)
Also performance improvements on heavy used systems unlocks:
Wait what? I don't get why performance improvement implies reliability and incident improvement.
For example, doing dangerous thing might be faster (no bound checks, weaker consistency guarantee, etc), but it clearly tend to be a reliability regression.
First, if a performance optimization is a reliability regression, it was done wrong. A bounds check is removed because something somewhere else is supposed to already guaratee it won't be violated, not just in a vacuum. If the guarantee stands, removing the extra check makes your program faster and there is no reliability regression whatsoever.
And how does performance improve reliability? Well, a more performant service is harder to overwhelm with a flood of requests.
It does not need to be an explicit check (i.e. a condition checking that your index is not out of bounds). You may structure your code in such a way that it becomes a mathematical impossibility to exceed the bounds. For a dumb trivial example, you have an array of 500 bytes and are accessing it with an 8-bit unsigned index - there's no explicit bounds check, but you can never exceed its bounds, because the index may only be 0-255.
Of course this is a very artificial and almost nonsensical example, but that is how you optimize bounds checks away - you just make it impossible for the bounds to be exceeded through means other than explicitly checking.
Which is fine, but the vast majority of the things that get presented aren’t bothering to benchmark against my use (for a whole lotta mes). They come from someone scratching an itch and solving it for a target audience of one and then extrapolating and bolting on some benchmarks. And at the sizes you’re talking about, how many tooling authors have the computing power on hand to test that?
Well obviously that would happen mostly only on the biggest business scales or maybe academic research; one example from Nvidia, which showcases Apache Spark with GPU acceleration to process "tens of terabytes of JSON data":
Who is the target audience? I truly wonder who will process TB-sized data using jq? Either it's in a database already, in which case you're using the database to process the data, or you're putting it in a database.
Either way, I have really big doubts that there will be ever a significant amount of people who'd choose jq for that.
There was a thread yesterday where a company rewrote a similar JSON processing library in Go because they were spending $100,000s on serving costs using it to filter vast amounts of data: https://news.ycombinator.com/item?id=47536712
I absolutely understand what you're saying. It makes complete sense. But I will never, ever shake the sense that software that isn't as fast as possible is offensive, immoral, delinquent -- the result of sloth, lassitude, lack of imagination, and a general hostility toward our noble Art.
"Fast enough" will always bug me. "Still ahead of network latency" will always sound like the dog ate your homework. I understand the perils of premature optimization, but not a refusal to optimize.
I get the sentiment, but everybody thinks that, and in aggregate, you get death by a thousand paper cuts.
It’s the same sentiment as “Individuals don’t matter, look at how tiny my contribution is.”. Society is made up of individuals, so everybody has to do their part.
> 9/10 whatever tooling you are using now will be perfectly fine.
It is not though. Software is getting slower faster than hardware is getting quicker. We have computers that are easily 3–4+ orders of magnitudes faster than what we had 40 years ago, yet everything has somehow gotten slower.
For better or worse, Claude is my intuitive interface to jq. I don't use it frequently, and before I would have to look up the commands every time, and slowly iterate it down to what I needed.
The syntax makes perfect sense when you understand the semantics of the language.
Out of curiosity, have you read the jq manpage? The first 500 words explain more or less the entire language and how it works. Not the syntax or the functions, but what the language itself is/does. The rest follows fairly easily from that.
Was about to post exactly this... It is impressive engineering wise, but for data and syntax, ease of use or all the great features, I care about that more. Speed isn't that important to me for a lot of these tools.
If I/you was working with JSON of that size where this was important, id say you probably need to stop using JSON! and some other binary or structured format... so long as it has some kinda tooling support.
And further if you are doing important stuff in the CLI needing a big chain of commands, you probably should be programming something to do it anyways...
that's even before we get to the whole JSON isn't really a good data format whatsoever... and there are many better ways. The old ways or the new ways. One day I will get to use my XSLT skills again :D
I agree for some things, but not for tools or "micro-software" like jq that can get called a LOT in an automated process. Every order of magnitude saved for the latter category can be meaningful.
Maybe look at it from another perspective.
Better performance == less CPU cycles wasted. Consider how many people use jq daily and think about how much energy could be saved by faster implementations.
In times like this where energy is becoming more scarce we should think about things like this.
> Consider how many people use jq daily and think about how much energy could be saved by faster implementations.
Say a number; make a real argument. Don't just wave your hand and say "just imagine how right I could be about this vague notion if we only knew the facts"
> Now what I'd really want is a jq that's more intuitive and easier to understand
Unfortunately I don’t recall the name, but there was something submitted to HN not too long ago (I think it was still 2026) which was like jq but used JavaScript syntax.
> I see the same thing repeated in various front end tooling too. They all claim to be _much_ faster than their counterpart.
>
> 9/10 whatever tooling you are using now will be perfectly fine
Are you working in frontend? On non-trivial webapps? Because this is entirely wrong in my experience. Performance issues are the #1 complaint of everyone on the frontend team. Be that in compiling, testing or (to a lesser extend) the actual app.
Worked on front end for years. Rarely ever hear people talking about performance issues. I was among the very few people who knew how to use the dev tools to investigate memory leak or heard of memlab.
Either the team I worked at was horrible, or you are from Google/Meta/Walmart where either everyone is smart or frondend performance is directly related to $$.
Uh, I've worked for a few years as a frontend dev, as in literal frontend dev - at that job my responsibility started at consuming and ended at feeding backend APIs, essentially.
From that I completely agree with your statement - however, you're not addressing the point he makes which kinda makes your statement completely unrelated to his point
99.99% of all performance issues in the frontend are caused by devs doing dumb shit at this point
The frameworks performance benefits are not going to meaningfully impact this issue anymore, hence no matter how performant yours is, that's still going to be their primary complaint across almost all complex rwcs
And the other issue is that we've decided that complex transpiling is the way to go in the frontend (typescript) - without that, all built time issues would magically go away too. But I guess that's another story.
It was a different story back when eg meteorjs was the default, but nowadays they're all fast enough to not be the source of the performance issues
> The vast majority of Linux kernel performance improvement patches probably have way less of a real world impact than this.
unlikely given that the number they are multiplying by every improvement is far higher than "times jq is run in some pipeline". Even 0.1% improvement in kernel is probably far far higher impact than this
I wonder so often about many new CLI tools whose primary selling point is their speed over other tools. Yet I personally have not encountered any case where a tool like jq feels incredibly slow, and I would feel the urge to find something else.
What do people do all day that existing tools are no longer enough? Or is it that kind of "my new terminal opens 107ms faster now, and I don't notice it, but I simply feel better because I know"?
I process TB-size ndjson files. I want to use jq to do some simple transformations between stages of the processing pipeline (e.g. rename a field), but it so slow that I write a single-use node or rust script instead.
I would love, _love_ to know more about your data formats, your tools, what the JSON looks like, basically as much as you're willing to share. :)
For about a month now I've been working on a suite of tools for dealing with JSON specifically written for the imagined audience of "for people who like CLIs or TUIs and have to deal with PILES AND PILES of JSON and care deeply about performance".
For me, I've been writing them just because it's an "itch". I like writing high performance/efficient software, and there's a few gaps that it bugged me they existed, that I knew I could fill.
I'm having fun and will be happy when I finish, regardless, but it would be so cool if it happened to solve a problem for someone else.
> The query language is deliberately less expressive than jq's. jsongrep is a search tool, not a transformation tool-- it finds values but doesn't compute new ones. There are no filters, no arithmetic, no string interpolation.
Mind me asking what sorts of TB json files you work with? Seems excessively immense.
Command-line Tools can be 235x Faster than your Hadoop Cluster (2014)
Conclusion: Hopefully this has illustrated some points about using and abusing tools like Hadoop for data processing tasks that can better be accomplished on a single machine with simple shell commands and tools.
JQ is very convenient, even if your files are more than 100GB.
I often need to extract one field from huge JSON line files, I just pipe jq to it to get results. It's slower, but implementing proper data processing will take more time.
> Now I'm really curious. What field are you in that ndjson files of that size are common?
I'm not OP,but structured JSON logs can easily result in humongous ndjson files, even with a modest fleet of servers over a not-very-long period of time.
Replying here because the other comment is too deeply nested to reply.
Even if it's once off, some people handle a lot of once-offs, that's exactly where you need good CLI tooling to support it.
Sure jq isn't exactly super slow, but I also have avoided it in pipelines where I just need faster throughput.
rg was insanely useful in a project I once got where they had about 5GB of source files, a lot of them auto-generated. And you needed to find stuff in there. People were using Notepad++ and waiting minutes for a query to find something in the haystack. rg returned results in seconds.
The use case could be e.g. exactly processing an old trove of logs into something more easily indexed and queryable, and you might want to use jq as part of that processing pipeline
Fair, but for a once-off thing performance isn't usually a major factor.
The comment I was replying to implied this was something more regular.
EDIT: why is this being downvoted? I didn't think I was rude. The person I responded to made a good point, I was just clarifying that it wasn't quite the situation I was asking about.
At scale, low performance can very easily mean "longer than the lifetime of the universe to execute." The question isn't how quickly something will get done, but whether it can be done at all.
Good point. I said it above, but I'll repeat it here that I shouldn't have discounted how frequent once offs can be. I've worked in support before so I really should've known better
Certain people/businesses deal with one-off things every day. Even for something truly one-off, if one tool is too slow it might still be the difference between being able to do it once or not at all.
If you work at a hyperscaler, service log volume borders on the insane, and while there is a whole pile of tooling around logs, often there's no real substitute for pulling a couple of terabytes locally and going to town on them.
> often there's no real substitute for pulling a couple of terabytes locally and going to town on them.
Fully agree. I already know the locations of the logs on-disk, and ripgrep - or at worst, grep with LC_ALL=C - is much, much faster than any aggregation tool.
If I need to compare different machines, or do complex projections, then sure, external tooling is probably easier. But for the case of “I know roughly when a problem occurred / a text pattern to match,” reading the local file is faster.
We parse JSON responses for dashboards, alerting, etc. Thousands of nodes, depending on the resolution of your monitoring you could see improvements here.
- They take Rust for performance or FAVORITE_LANG for credentials
- Claude implements small subset of features
- Benchmark subset
- Claim win, profit on showcase
Note: this particular project doesn't have many visible tells, but there's pattern of overdocumentation (17% comment-to-code ratio, >1000 words in README, Claude-like comment patterns), so it might be a guided process.
I still think that the project follows the "subset is faster than set" trend.
You don't know something is slow until you encounter a use case where the speed becomes noticeable. Then you see the slowness across the board. If you can notice that a command hasn't completed and you are able to fully process a thought about it, it's slow(er than your mind, ergo slow!).
Usually, a perceptive user/technical mind is able to tweak their usage of the tools around their limitations, but if you can find a tool that doesn't have those limitations, it feels far more superior.
The only place where ripgrep hasn't seeped into my workflow for example, is after the pipe and that's just out of (bad?) habit. So much so, sometimes I'll do this foolishly rg "<term>" | grep <second filter>; then proceed to do a metaphoric facepalm on my mind. Let's see if jg can make me go jg <term> | jq <transformation> :)
Well grep is just better sometimes. Like you want to copy some lines and grep at the end of a pipeline is just easier than rg -N to suppress line numbers. Whatever works, no need to facepalm.
Prioritizing SEO-ing speed over supporting the same features/syntax (especially without an immediately prominent disclosure of these deficiencies) = marketing bullshit
A faster jq except it can't do what jq does... maybe I can use this as a pre-filter when necessary.
I am not sure if it was simon or pg who might've quoted this but I remembered a quote about that a 2 magnitude order in speed (quantity) is a huge qualititative change in it of itself.
If somebody needs performance, they probably shouldn't be calling out to a separate process for json of all things, no?
(Honestly, who even still writes shell scripts? Have a coding agent write the thing in a real scripting language at least, they aren't phased by the boilerplate of constructing pipelines with python or whatever. I haven't written a shell script in over a year now.)
If you’re writing the script to be used by multiple people, or on multiple systems, or for CI runners, or in containers, etc. then there’s no guarantee of having Python (mostly for the container situation, but still), much less of its version. It’s far too easy to accidentally use a feature or syntax that you took for granted, because who would still be using 3.7 today, anyway? I say this from painful recent experience.
Plus, for any script that’s going to be fetching or posting anything over a network, the LLM will almost certainly want to include requests, so now you either have to deal with dependencies, or make it use urllib.
In contrast, there’s an extremely high likelihood of the environment having a POSIX-compatible interpreter, so as long as you don’t use bash-isms (or zsh-isms, etc.), the script will probably work. For network access, the odds of it having curl are also quite high, moreso (especially in containers) than Python.
If you're distributing the script to other people then the benifit of using python and getting stuff like high quality argument parsing for free is even greater.
For people chewing through 50GB logs or piping JSON through cron jobs all day, a 2x speedup is measurable in wall time and cloud bill, not just terminal-brain nonsense. Most people won't care.
If jq is something you run a few times by hand, a "faster jq" is about as compelling as a faster toaster. A lot of these tools still get traction because speed is an easy pitch, and because some team hit one ugly bottleneck in CI or a data pipeline and decided the old tool was now unacceptable.
I deal with a fair amount of newline-delimited JSON in my day job, where each line in the file is a complete JSON object. I've seen this referred to as "jsonl", and it's not entirely uncommon for logs and other kinds of time-series data dumps. Do any of the popular JSON CLI tools work with this format? I didn't see any mention of it here.
Having used `jq` and `yq` (which followed from the former, in spirit), I have never had to complain about performance of the _latter_ which an order of magnitude (or several) _slower_ than the former. So if there's something faster than `jq`, it's laudable that the author of the faster tool accomplished such a goal, but in the broader context I'd say the performance benefit would be required by a niche slice of the userbase. People who analyse JSON-formatted logs, perhaps? Then again, newline-delimited JSON reigns supreme in that particular kind of scenario, making the point of a faster `jq` moot again.
However, as someone who always loved faster software and being an optimisation nerd, hat's off!
Integrating with server software, the performance is nice to have, as you can have say 100 kRPS requests coming in that need some jq-like logic. For CLI tool, like you said, the performance of any of them is ok, for most of the cases.
I keep an eye on jaq, but there are some holes in the story. jaq 3.0 is faster than Linux distro builds of jq, but jq built correctly is faster than jaq. As far as I can tell the performance reputation of jq is caused by bad distro packaging.
When initially opening the page it had broken colors in light mode. For anyone else encountering it: switch to dark mode and then back to light mode to fix it.
I would not be surprised at all if it's vibe coded. I have seen exactly the same thing myself.
I gave instruction to Claude to add a toggle button to a website where the value needs to be stored in local storage.
It is a very straightforward change. Just follow exactly how it is done for a different boolean setting and you are set. An intern can do that on the first day of their job.
Everything is done properly except that on page load, the stored setting is not read.
Which can be easily discovered if the author, with or without AI tools, has a test or manually goes through the entire workflow just once. I discovered the problem myself and fixed it.
Setting all of that aside -- even if this is not AI coded, at the least it shows the site owner doesn't have the basic care for its visitors to go through this important workflow to check if everything works properly.
And who cares if it's vibe-coded or not. Since when do we care more on the how than on the what? Are people looking at how a tool was coded before using it, as if it would accelerate confidence?
I learned a number of data processing cli tools: jq, mlr, htmlq, xsv, yq, etc; to name a few. Not to the level of completing advent of code or anything, but good enough for my day to day usage. It was never ending with the amount of formats I needed to extract data from, and the different syntax's. All that changed when I found nushell though, its replaced all of these tools for me. One syntax for everything, breath of fresh air!
+1. I switched to using Nushell as my daily driver around mid-2023 (0.84.0?) and use it in preference to other interactive tools. I do keep at hand jq, yq, and mlr because I need to exchange stuff with colleagues who don't use Nu.
Had to spend some efforts to set up completions, also there some small rough edges around commands discoverability, but anyway, much better than the previous oh-my-zsh setup
Ideally, wish it also had a flag to enforce users to write type annotations + compiling scripts as static binaries + a TUI library, and then I'd seriously consider it for writing small apps, but I like and appreciate it in the current state already
Same here, nushell is awesome! It helped me to automate so many more things than I did with any other shell. The syntax is so much more intuitive and coherent, which really helps a lot for someone who always forgot how to write ifs or loops in bash ^^
Second, some comments on the presentation: the horizontal violin graphs are nice, but all tools have the same colours, and so it's just hard to even spot where jsongrep is. I'd recommend grouping by tool and colour coding it. Besides, jq itself isn't in the graphs at all (but the title of the post made me think it would be!).
Last, xLarge is a 190MiB file. I was surprised by that. It seems too low for xLarge. I daily check 400MiB json documents, and sometimes GiB ones.
Hey thank you! OP here, yes I was struggling to find large enough documents to run the benchmarks on, the range currently on the benchmark data is ~106 B - ~190MB, which I think covers the majority of quick task workloads, but would love to have large documents, if there's an public ones you can thinking of I'd like to know!
> Jq is a powerful tool, but its imperative filter syntax can be verbose for common path-matching tasks. jsongrep is declarative: you describe the shape of the paths you want, and the engine finds them.
IMO, this isn't a common use case. The comparison here is essentially like Java vs Python. Jq is perfectly fine for quick peeking. If you actually need better performance, there are always faster ways to parse JSON than using a CLI.
The data viz of the benchmarks is really rough. I think you’d get a lot of leverage out of rebuilding it and using colors and/or shapes to extract additional dimensions. Nobody wants to scan through raw file paths as labels to try and figure out what the hell the results are
Having the equivalent jq expression in these examples might help to compare expressiveness, and it might help me see if jq could “just” use a DFA when a (sub)query admits one. grep, ripgrep, etc change algorithms based on the query and that makes the speed improvements automatic.
Also "jg" reads very similar to "jq", and initially I thought he was talking about "jq" all along, and I was like: where can I see the "jasongrep" examples? Threw me off for a minute.
Quick question:
Isn't the construction of a NFA - DFA a O(2^n) algorithm? If a JSON file has a couple hundred values, its equivalent NFA will have a similar amount, and the DFA will have 2^100 states, so I must be missing something.
theory is one thing but the cpu cache is the real bottleneck here... here is a small visual breakdown of how these arrays look in memory and why pointer chasing is so expensive compared to the actual logic: https://vectree.io/c/json-array-memory-indexing
basically the double jump to find values in the heap is what slows down these tools most
I can see that in practice the bottleneck isn't the automata construction, I'm just curious of how the construction is approached with such a super-exponential conversion algorithm
One problem I have not seen addressed by jq or alterataives, perhaps this one addresses it, is "JSON-like" data. That is, JSON that is not contained in a JSON file
For example, web pages sometimes contain inline "JSON". But as this is not a proper JSON file, jq-style utilties cannot process it
The solution I have used for years is a simple utility written in C using flex^1 (a "filter") that reformats "JSON" on stdin, regardless of whether the input is a proper JSON file or not, into stdout that is line-delimited, human-readable and therefore easy to process with common UNIX utilities
The size of the JSON input does not affect the filter's memory usage. Generally, a large JSON file is processed at the same speed with the same resource usage as a small one
The author here has provided musl static-pie binaries instead of glibc. HN commenters seeking to discredit musl often claim glibc is faster
I am excited for some alternative syntax to jq's. I haven't given much thought to how I'd write a new JSON query syntax if I were writing things from scratch, but I personally never found the jq syntax intuitive. Perhaps I haven't given it enough effort to learn properly.
You don't learn it properly. It's not supposed to be intuitive, it's supposed to be concise at the cost of it being intuitive. Would be like somebody saying typing words in to Google is more intuitive than writing regex.
jq is supposed to fit in to other bash scripts as a one liner. That's it's super power. I know very few people who write regex on the fly either (unless you were using it everyday) they check the documentation and flesh it out when they need it.
Just use Claude to generate the jq expression you need and test it.
I’d install it via cargo anyway and that would build it for arm64.
If the arm64 version was on homebrew (didn’t check if it is but assume not because it’s not mentioned on the page), I’d install it from there rather than from cargo.
I don’t really manually install binaries from GitHub, but it’s nice that the author provides binaries for several platforms for people that do like to install it that way.
Really? That is your response? This is an high quality article from someone who spend a lot of time implementing a cool tool and also sharing the intricate inner workings of it. And your response is, "eh there are no official binaries for my platform". Give them some credit! Be a little more constructive!
Since the query compilation needs exponential time, I wonder how large the queries can be before jsongrep becomes slower than all the other tools. In that regard, I think the library could benefit from some functionality for query compilation at compile-time.
Thank you. Very cool. Going to try embedding this into my JSON viewer. One thing I’ve struggled with is that live querying in the UI is constrained by performance.
forgive me my rant, but when I see "just install it with cargo" I immediately lose interest. How many GB do I have to install just to test a little tool? sorry, not gonna do that
I was a bit skeptical at first, but after reading more into jsongrep, it's actually very good. Only did a very quick test just now, and after stumbling over slightly different syntax to jq, am actually quite impressed. Give it a try
What were your syntax stumbling blocks? I must be honest I've used jq enough but can never remember the syntax. It's one of the worst things about jq IMO (not the speed, even though I'm a fan of speedups). There's something ungrokkable about that syntax for me.
Some bits of the site are hard to read "takes a query and a JSON input" query is in white and the background of the site is very light which makes it hard to read.
If the author cares, I can’t read everything on this page. The command snippets have a “BASH” pill in the top left that covers up the command I’m supposed to run. And then there are, I guess topic headings or something that are white-on-white text, so honestly I don’t know what they say or what they are.
Another alternative is oj, https://github.com/ohler55/ojg. I don't know how the performance compares to jq or any others but it does use JSONPath as the query language. It has a few other options for making nicely formatted JSON and colorizing JSON.
Many Useless Uses of cat in this documentation. You never need to do `cat file | foo`, you can just do `<file foo`. cat is for concatenating inputs, you never need it for a single input.
As someone who worked with Unix/Linux and command line arguments for 30 years and still "abuse" cat like the documentation, I regularly hear this complaint.
Yes, "cmd <file" is more efficient for the computer but not for the reader in many cases.
I read from left to the right and the pipeline might be long or "cmd" might have plenty of arguments (or both).
Having "cat file | cmd" immediately gives me the context for what I am working with and corresponds well with "take this file, do this, then that, etc" with it) and makes it easier for me to grok what is happening (the first operation will have some kind of input from stdin).
Without that, the context starts with the (first) operation like in the sentence "do this operation, on this file (,then this, etc)". I might not be familiar with it or knowing the arguments it expects.
At least for me, the first variant comes more naturally and is quicker to follow (in most cases), so unless it is performance sensitive that is what I end up with (and cat is insanely fast for most cases).
Jq's syntax is so arcane I can never remember it and always need to look up how to get a value from simple JSON.
That’s interesting! Can you say a little more? I find jq’s syntax and semantics to be simple and intuitive. It’s mostly dots, pipes, and brackets. It’s a lot like writing shell pipelines imo. And I tend to use it in the same way. Lots of one-time use invocations, so I spend more time writing jq filters than I spend reading them.
I suspect my use cases are less complex than yours. Or maybe jq just fits the way I think for some reason.
I dream of a world in which all CLI tools produce and consume JSON and we use jq to glue them together. Sounds like that would be a nightmare for you.
I'm not GP, I use jq all the time, but I each time I use it I feel like I'm still a beginner because I don't get where I want to go on the first several attempts. Great tool, but IMO it is more intuitive to JSON people that want a CLI tool than CLI people that want a JSON tool. In other words, I have my own preconceptions about how piping should work on the whole thing, not iterating, and it always trips me up.
Here's an example of my white whale, converting JSON arrays to TSV.
cat input.json | jq -S '(first|keys | map({key: ., value: .}) | from_entries), (.[])' | jq -r '[.[]] | @tsv' > out.tsv
oh my god how could I have been doing this for so long and not realize that you can redirect before your binary.
I knew cat was an anti-pattern, but I always thought it was so unreadable to redirect at the end
Here's an easier to understand query for what you're trying to do (at least it's easier to understand for me):
That whole map and from entries throws it off. It's not a good use for what you're doing. tsv expects a bunch of arrays, whereas you're getting a bunch of objects (with the header also being one) and then converting them to arrays. That is an unnecessary step and makes it a little harder to understand.Thanks for sharing, this is much better, though I actually think it is the perfect example to explain something that is brain-slippery about jq
look at $cols | $cols
my brain says hmm that's a typo, clearly they meant ; instead of | because nothing is getting piped, we just have two separate statements. Surely the assignment "exhausts the pipeline" and we're only passing null downstream
the pipelining has some implicit contextual stuff going on that I have to arrive at by trial and error each time since it doesn't fit in my worldview while I'm doing other shell stuff
Honestly both of those make me do the confused-dog-head-tilt thing. I'd go for something sexp based, perhaps with infix composition, map, and flatmap operators as sugar.
I find it much harder to remember / use each time then awk
Sound similar to how power shell works, and it’s not great. Plain text is better.
I think the big problem is it's a tool you usually reach for so rarely you never quite get the opportunity to really learn it well, so it always remains in that valley of despair where you know you should use it, but it's never intuitive or easy to use.
It's not unique in that regard. 'sed' is Turing complete[1][2], but few people get farther than learning how to do a basic regex substitution.
[1] https://catonmat.net/proof-that-sed-is-turing-complete
[1] And arguably a Turing tarpit.
I was just going to say, jq is like sed in that I only use 1% of it 99% of the time, but unlike sed in that I'm not aware of any clearly better if less ubiquitous alternatives to the 1% (e.g., Perl or ripgrep for simple regex substitutions in pipelines because better regex dialects).
Closest I've come, if you're willing to overlook its verbosity and (lack of) speed, is actually PowerShell, if only because it's a bit nicer than Python or JavaScript for interactive use.
Shameless plug, but you might like this: https://github.com/IvanIsCoding/celq
jq is the CLI I like the most, but sometimes even I struggled to understand the queries I wrote in the past. celq uses a more familiar language (CEL)
Cool tool! Really appreciate the shoutout to gron in the readme, thanks! :)
I had never heard of CEL, looks useful though, thanks for posting this!
CEL looks interesting and useful, though it isn't common nor familiar imo (not for me at least). Quoting from https://github.com/google/cel-spec
That’s some fair criticism, but the same page tells that the language wanted to have a similar syntax to C and JavaScript.
I think my personal preference for syntax would be Python’s. One day I want to try writing a query tool with https://github.com/pydantic/monty
Funny that everyone is linking the tools they wrote for themselves to deal with this problem. I am no exception. I wrote one that just lets you write JavaScript. Imagine my surprise that this extremely naive implementation was faster than jq, even on large files.
https://crespo.business/posts/dq-its-just-js/Like I did with regex some years earlier, I worked on a project for a few weeks that required constant interactions with jq, and through that I managed to lock in the general shape of queries so that my google hints became much faster.
Of course, this doesn't matter now, I just ask an LLM to make the query for me if it's so complex that I can't do it by hand within seconds.
To fix this I recently made myself a tiny tool I called jtree that recursively walks json, spitting out one line per leaf. Each line is the jq selector and leaf value separated by "=".
No more fiddling around trying to figure out the damn selector by trying to track the indentation level across a huge file. Also easy to pipe into fzf, then split on "=", trim, then pass to jq
You might like https://github.com/tomnomnom/gron
I completely agree. I much prefer leveraging actual javascript to get what I need instead of spending time trying to fumble my way through jq syntax.
Check this out: https://crespo.business/posts/dq-its-just-js/
You don't have to use my implementation, you could easily write your own.
Highly recommend gron. https://github.com/tomnomnom/gron
or https://github.com/adamritter/fastgron
If we're plugging jq alternatives, I'll plug my own: https://git.sr.ht/~charles/rq
I was working at lot with Rego (the DSL for Open Policy Agent) and realized it was actually a pretty nice syntax for jq type use cases.
I agree, even trivial tasks require us to go back to jq's manual to learn how to write their language.
this and other reasons is why I built: https://github.com/dhuan/dop
I just ask Opus to generate the queries for me these days.
JMESPath is what I wish jq was. Consistent grammar. It only issue is it lacks the ability to convert JSON to other formats like CSV.
LOL ... I can absolutely feel your pain. That's exactly why I created for myself a graphical approach. I shared the first version with friends and it turned into "ColumnLens" (ImGUI on Mac) app. Here is a use case from the healthcare industry: https://columnlens.com/industries/medical
I also genuinely hate using jq. It is one of the only things that I rely heavily on AI.
You should try nushell or PowerShell which have built ins to convert json to objects. It makes it so easy.
Second this. Working with nushell is a joy.
At that point why don't we ask the AI directly to filter through our data? The AI query language is much more powerful.
Because the output you get can have hallucinations, which don’t happen with a deterministic tool. Furthermore, by getting the `jq` command you get something which is reusable, fast, offline, local, doesn’t send your data to a third-party, doesn’t waste a bunch of tokens, … Using an LLM to filter the data is worse in every metric.
I get that AI isn’t deterministic by definition, but IMHO it’s become the go-to response for a reason to not use AI, regardless of the use case.
I’ve never seen AI “hallucinate” on basic data transformation tasks. If you tell it to convert JSON to YAML, that’s what you’re going to get. Most LLMs are probably using something like jq to do the conversion in the background anyway.
AI experts say AI models don’t hallucinate, they confabulate.
Just because you haven't seen it hallucinate on these tasks doesn't mean it can't.
When I'm deciding what tool to use, my question is "does this need AI?", not "could AI solve this?" There's plenty of cases where its hard to write a deterministic script to do something, but if there is a deterministic option, why would you choose something that might give you the wrong answer? It's also more expensive.
The jq script or other script that an LLM generates is way easier to spot check than the output if you ask it to transform the data directly, and you can reuse it.
You can use a local LLM and you can ask it to use tools so it is faster.
"so it is faster" than what? A cloud hosted LLM? That's a pretty low bar. It's certainly not faster than jq.
There is hardware that is able to run jq but no a local AI model that's powerful enough to make the filtering reliable. Ex a raspberry pi
Because the input might be sensitive.
Because the input might be huge.
Because there is a risk of getting hallucinations in the output.
Isn't this obvious?
...and because it's going to burn a million times the energy of what jq would require.
You really need to go and learn about the concept of determinism and why for some tasks we need and want deterministic solutions.
It's an important idea in computer science. Go and learn.
You need to learn to adapt to the real world where most things are not deterministic. Go and learn.
I already know that. That's why we have deterministic algorithms, to simplify that complexity. You have much to learn, witty answers mean nothing here, particularly empty witty answers, which are no better than jokes. Maybe stand-up comedy is your call in life.
That may be true, but do you not want determinism where possible, especially within this context, i.e. filtering data?
Is your argument that the world isn't deterministic and so we should also apply nondeterminism to filtering json data?
yeah I literally just use gemini / claude to one-shot JQ queries now
I appreciate performance as much as the next person; but I see this endless battle to measure things in ns/us/ms as performative.
Sure there are 0.000001% edge cases where that MIGHT be the next big bottleneck.
I see the same thing repeated in various front end tooling too. They all claim to be _much_ faster than their counterpart.
9/10 whatever tooling you are using now will be perfectly fine. Example; I use grep a lot in an ad hoc manner on really large files I switch to rg. But that is only in the handful of cases.
Whenever you have this kind of impressions on some development, here are my 2 cents: just think "I'm not the target audience". And that's fine.
The difference between 2ms and 0.2ms might sound unneeded, or even silly to you. But somebody, somewhere, is doing stream processing of TB-sized JSON objects, and they will care. These news are for them.
I remember when I was coming up on the command line and I'd browse the forums at unix.com. Someone would ask how to do a thing and CFAJohnson would come in with a far less readable solution that was more performative (probably replacing calls to external tools with Bash internals, but I didn't know enough then to speak intelligently about it now).
People would say, "Why use this when it's harder to read and only saves N ms?" He'd reply that you'd care about those ms when you had to read a database from 500 remote servers (I'm paraphrasing. He probably had a much better example.)
Turns out, he wrote a book that I later purchased. It appears to have been taken over by a different author, but the first release was all him and I bought it immediately when I recognized the name / unix.com handle. Though it was over my head when I first bought it, I later learned enough to love it. I hope he's on HN and knows that someone loved his posts / book.
https://www.amazon.com/Pro-Bash-Programming-Scripting-Expert...
Wow that takes me back. I used to lurk on unix.com when I was starting with bash and perl and would see CFAJohnson's terse one-liners all the time. I enjoyed trying my own approaches to compare performance, conciseness and readability - mainly for learning. Some of the awk stuff was quite illuminating in my understanding of how powerful awk could be. I remember trying different approaches to process large files at first with awk and then with Perl. Then we discovered Oracle's external tables which turned out to be clear winner. We have a lot more options now with fantastic performance.
Also as someone who looks at latency charts too much, what happens is a request does a lot in series and any little ms you can knock off adds up. You save 10ms by saving 10 x 1ms. And if you are a proxyish service then you are a 10ms in a chain that might be taking 200 or 300ms. It is like saving money, you have to like cut lots of small expenses to make an impact. (unless you move etc. but once you done that it is small numerous things thay add up)
Also performance improvements on heavy used systems unlocks:
Cost savings
Stability
Higher reliability
Higher throughput
Fewer incidents
Lower scaling out requirements.
Wait what? I don't get why performance improvement implies reliability and incident improvement.
For example, doing dangerous thing might be faster (no bound checks, weaker consistency guarantee, etc), but it clearly tend to be a reliability regression.
First, if a performance optimization is a reliability regression, it was done wrong. A bounds check is removed because something somewhere else is supposed to already guaratee it won't be violated, not just in a vacuum. If the guarantee stands, removing the extra check makes your program faster and there is no reliability regression whatsoever.
And how does performance improve reliability? Well, a more performant service is harder to overwhelm with a flood of requests.
"Removing an extra check", so there is a check, so the check is not removed?
It does not need to be an explicit check (i.e. a condition checking that your index is not out of bounds). You may structure your code in such a way that it becomes a mathematical impossibility to exceed the bounds. For a dumb trivial example, you have an array of 500 bytes and are accessing it with an 8-bit unsigned index - there's no explicit bounds check, but you can never exceed its bounds, because the index may only be 0-255.
Of course this is a very artificial and almost nonsensical example, but that is how you optimize bounds checks away - you just make it impossible for the bounds to be exceeded through means other than explicitly checking.
Less OOMs, less timeouts, less noisy neighbors problems affecting other apps
Which is fine, but the vast majority of the things that get presented aren’t bothering to benchmark against my use (for a whole lotta mes). They come from someone scratching an itch and solving it for a target audience of one and then extrapolating and bolting on some benchmarks. And at the sizes you’re talking about, how many tooling authors have the computing power on hand to test that?
> "somebody, somewhere, is doing stream processing of TB-sized JSON objects"
That's crazy to think about. My JSON files can be measured in bytes. :-D
Well obviously that would happen mostly only on the biggest business scales or maybe academic research; one example from Nvidia, which showcases Apache Spark with GPU acceleration to process "tens of terabytes of JSON data":
https://developer.nvidia.com/blog/accelerating-json-processi...
All files can be measured in bytes. :)
You, sir or ma'am, are a first class smarty pants.
But even in this example, the 2ms vs 0.2 is irrelevant - its whatever the timings are for TB-size objects.
So went not compare that case directly? We'd also want to see the performance of the assumed overheads i.e. how it scales.
Who is the target audience? I truly wonder who will process TB-sized data using jq? Either it's in a database already, in which case you're using the database to process the data, or you're putting it in a database.
Either way, I have really big doubts that there will be ever a significant amount of people who'd choose jq for that.
There was a thread yesterday where a company rewrote a similar JSON processing library in Go because they were spending $100,000s on serving costs using it to filter vast amounts of data: https://news.ycombinator.com/item?id=47536712
I absolutely understand what you're saying. It makes complete sense. But I will never, ever shake the sense that software that isn't as fast as possible is offensive, immoral, delinquent -- the result of sloth, lassitude, lack of imagination, and a general hostility toward our noble Art.
"Fast enough" will always bug me. "Still ahead of network latency" will always sound like the dog ate your homework. I understand the perils of premature optimization, but not a refusal to optimize.
And I doubt I'm alone.
I get the sentiment, but everybody thinks that, and in aggregate, you get death by a thousand paper cuts.
It’s the same sentiment as “Individuals don’t matter, look at how tiny my contribution is.”. Society is made up of individuals, so everybody has to do their part.
> 9/10 whatever tooling you are using now will be perfectly fine.
It is not though. Software is getting slower faster than hardware is getting quicker. We have computers that are easily 3–4+ orders of magnitudes faster than what we had 40 years ago, yet everything has somehow gotten slower.
True. I feel like the main way a tool could differentiate from jq is having more intuitive syntax and many real world examples to show off the syntax.
For better or worse, Claude is my intuitive interface to jq. I don't use it frequently, and before I would have to look up the commands every time, and slowly iterate it down to what I needed.
The syntax makes perfect sense when you understand the semantics of the language.
Out of curiosity, have you read the jq manpage? The first 500 words explain more or less the entire language and how it works. Not the syntax or the functions, but what the language itself is/does. The rest follows fairly easily from that.
Was about to post exactly this... It is impressive engineering wise, but for data and syntax, ease of use or all the great features, I care about that more. Speed isn't that important to me for a lot of these tools.
If I/you was working with JSON of that size where this was important, id say you probably need to stop using JSON! and some other binary or structured format... so long as it has some kinda tooling support.
And further if you are doing important stuff in the CLI needing a big chain of commands, you probably should be programming something to do it anyways...
that's even before we get to the whole JSON isn't really a good data format whatsoever... and there are many better ways. The old ways or the new ways. One day I will get to use my XSLT skills again :D
I agree for some things, but not for tools or "micro-software" like jq that can get called a LOT in an automated process. Every order of magnitude saved for the latter category can be meaningful.
Maybe look at it from another perspective. Better performance == less CPU cycles wasted. Consider how many people use jq daily and think about how much energy could be saved by faster implementations. In times like this where energy is becoming more scarce we should think about things like this.
I agree, but in this age of widespread LLM use, that's only marginal.
> Consider how many people use jq daily and think about how much energy could be saved by faster implementations.
Say a number; make a real argument. Don't just wave your hand and say "just imagine how right I could be about this vague notion if we only knew the facts"
Then this is for the handful of cases for you. When it matters it matters.
Yes
I don't think I remember one case where jq wasn't fast enough
Now what I'd really want is a jq that's more intuitive and easier to understand
> Now what I'd really want is a jq that's more intuitive and easier to understand
Unfortunately I don’t recall the name, but there was something submitted to HN not too long ago (I think it was still 2026) which was like jq but used JavaScript syntax.
> I see the same thing repeated in various front end tooling too. They all claim to be _much_ faster than their counterpart.
>
> 9/10 whatever tooling you are using now will be perfectly fine
Are you working in frontend? On non-trivial webapps? Because this is entirely wrong in my experience. Performance issues are the #1 complaint of everyone on the frontend team. Be that in compiling, testing or (to a lesser extend) the actual app.
Worked on front end for years. Rarely ever hear people talking about performance issues. I was among the very few people who knew how to use the dev tools to investigate memory leak or heard of memlab.
Either the team I worked at was horrible, or you are from Google/Meta/Walmart where either everyone is smart or frondend performance is directly related to $$.
"performance is directly related to $$"
It is. Company size is moot. See https://wpostats.com for starters.
There are some really fast tools out there for compiling FE these days, and that's probably to what they refer. Testing is still a slog.
Uh, I've worked for a few years as a frontend dev, as in literal frontend dev - at that job my responsibility started at consuming and ended at feeding backend APIs, essentially.
From that I completely agree with your statement - however, you're not addressing the point he makes which kinda makes your statement completely unrelated to his point
99.99% of all performance issues in the frontend are caused by devs doing dumb shit at this point
The frameworks performance benefits are not going to meaningfully impact this issue anymore, hence no matter how performant yours is, that's still going to be their primary complaint across almost all complex rwcs
And the other issue is that we've decided that complex transpiling is the way to go in the frontend (typescript) - without that, all built time issues would magically go away too. But I guess that's another story.
It was a different story back when eg meteorjs was the default, but nowadays they're all fast enough to not be the source of the performance issues
Fair, but agentic tooling can benefit quite a lot from this
Opencode, ClaudeCode, etc, feel slow. Whatever make them faster is a win :)
The 2ms it takes to run jq versus the 0.2ms to run an alternative is not why your coding agent feels slow.
Still, jq is run a whole lot more than it used to be due to coding agents, so every bit helps.
The vast majority of Linux kernel performance improvement patches probably have way less of a real world impact than this.
> The vast majority of Linux kernel performance improvement patches probably have way less of a real world impact than this.
unlikely given that the number they are multiplying by every improvement is far higher than "times jq is run in some pipeline". Even 0.1% improvement in kernel is probably far far higher impact than this
Jq is run a ton by AIs, and that is only increasing.
I can't take seriously any talk about performance if the tools are going to shell out. It's just not a bottleneck.
It's not running jq locally that's causing that
I wonder so often about many new CLI tools whose primary selling point is their speed over other tools. Yet I personally have not encountered any case where a tool like jq feels incredibly slow, and I would feel the urge to find something else. What do people do all day that existing tools are no longer enough? Or is it that kind of "my new terminal opens 107ms faster now, and I don't notice it, but I simply feel better because I know"?
I process TB-size ndjson files. I want to use jq to do some simple transformations between stages of the processing pipeline (e.g. rename a field), but it so slow that I write a single-use node or rust script instead.
I would love, _love_ to know more about your data formats, your tools, what the JSON looks like, basically as much as you're willing to share. :)
For about a month now I've been working on a suite of tools for dealing with JSON specifically written for the imagined audience of "for people who like CLIs or TUIs and have to deal with PILES AND PILES of JSON and care deeply about performance".
For me, I've been writing them just because it's an "itch". I like writing high performance/efficient software, and there's a few gaps that it bugged me they existed, that I knew I could fill.
I'm having fun and will be happy when I finish, regardless, but it would be so cool if it happened to solve a problem for someone else.
This reminds me of someone who wrote a regex tool that matches by compiling regexes (at runtime of the tool) via LLVM to native code.
You could probably do something similar for a faster jq.
This isn't for you then
> The query language is deliberately less expressive than jq's. jsongrep is a search tool, not a transformation tool-- it finds values but doesn't compute new ones. There are no filters, no arithmetic, no string interpolation.
Mind me asking what sorts of TB json files you work with? Seems excessively immense.
> Uses jq for TB json files
> Hadoop: bro
> Spark: bro
> hive: bro
> data team: bro
made me remember this article
<https://adamdrake.com/command-line-tools-can-be-235x-faster-...>
JQ is very convenient, even if your files are more than 100GB. I often need to extract one field from huge JSON line files, I just pipe jq to it to get results. It's slower, but implementing proper data processing will take more time.
are those tools known for their fast json parsers?
Now I'm really curious. What field are you in that ndjson files of that size are common?
I'm sure there are reasons against switching to something more efficient–we've all been there–I'm just surprised.
> Now I'm really curious. What field are you in that ndjson files of that size are common?
I'm not OP,but structured JSON logs can easily result in humongous ndjson files, even with a modest fleet of servers over a not-very-long period of time.
So what's the use case for keeping them in that format rather than something more easily indexed and queryable?
I'd probably just shove it all into Postgres, but even a multi terabyte SQLite database seems more reasonable.
Replying here because the other comment is too deeply nested to reply.
Even if it's once off, some people handle a lot of once-offs, that's exactly where you need good CLI tooling to support it.
Sure jq isn't exactly super slow, but I also have avoided it in pipelines where I just need faster throughput.
rg was insanely useful in a project I once got where they had about 5GB of source files, a lot of them auto-generated. And you needed to find stuff in there. People were using Notepad++ and waiting minutes for a query to find something in the haystack. rg returned results in seconds.
You make some good points. I've worked in support before, so I shouldn't have discounted how frequent "once-offs" can be.
The use case could be e.g. exactly processing an old trove of logs into something more easily indexed and queryable, and you might want to use jq as part of that processing pipeline
Fair, but for a once-off thing performance isn't usually a major factor.
The comment I was replying to implied this was something more regular.
EDIT: why is this being downvoted? I didn't think I was rude. The person I responded to made a good point, I was just clarifying that it wasn't quite the situation I was asking about.
At scale, low performance can very easily mean "longer than the lifetime of the universe to execute." The question isn't how quickly something will get done, but whether it can be done at all.
Good point. I said it above, but I'll repeat it here that I shouldn't have discounted how frequent once offs can be. I've worked in support before so I really should've known better
Certain people/businesses deal with one-off things every day. Even for something truly one-off, if one tool is too slow it might still be the difference between being able to do it once or not at all.
Deal with really big log files, mostly.
If you work at a hyperscaler, service log volume borders on the insane, and while there is a whole pile of tooling around logs, often there's no real substitute for pulling a couple of terabytes locally and going to town on them.
> often there's no real substitute for pulling a couple of terabytes locally and going to town on them.
Fully agree. I already know the locations of the logs on-disk, and ripgrep - or at worst, grep with LC_ALL=C - is much, much faster than any aggregation tool.
If I need to compare different machines, or do complex projections, then sure, external tooling is probably easier. But for the case of “I know roughly when a problem occurred / a text pattern to match,” reading the local file is faster.
We parse JSON responses for dashboards, alerting, etc. Thousands of nodes, depending on the resolution of your monitoring you could see improvements here.
It's a simple loop:
- Someone likes tool X
- Figures, that they can vibe code alternative
- They take Rust for performance or FAVORITE_LANG for credentials
- Claude implements small subset of features
- Benchmark subset
- Claim win, profit on showcase
Note: this particular project doesn't have many visible tells, but there's pattern of overdocumentation (17% comment-to-code ratio, >1000 words in README, Claude-like comment patterns), so it might be a guided process.
I still think that the project follows the "subset is faster than set" trend.
You don't know something is slow until you encounter a use case where the speed becomes noticeable. Then you see the slowness across the board. If you can notice that a command hasn't completed and you are able to fully process a thought about it, it's slow(er than your mind, ergo slow!).
Usually, a perceptive user/technical mind is able to tweak their usage of the tools around their limitations, but if you can find a tool that doesn't have those limitations, it feels far more superior.
The only place where ripgrep hasn't seeped into my workflow for example, is after the pipe and that's just out of (bad?) habit. So much so, sometimes I'll do this foolishly rg "<term>" | grep <second filter>; then proceed to do a metaphoric facepalm on my mind. Let's see if jg can make me go jg <term> | jq <transformation> :)
Well grep is just better sometimes. Like you want to copy some lines and grep at the end of a pipeline is just easier than rg -N to suppress line numbers. Whatever works, no need to facepalm.
Race between ripgrep and ugrep is entertaining.
Optimization = good
Prioritizing SEO-ing speed over supporting the same features/syntax (especially without an immediately prominent disclosure of these deficiencies) = marketing bullshit
A faster jq except it can't do what jq does... maybe I can use this as a pre-filter when necessary.
Speed is a quality in itself. We are so bugged down by slow stuff that we often ignore that and don’t actively search for another.
But every now and then a well-optimised tool/page comes along with instant feedback and is a real pleasure to use.
I think some people are more affected by that than others.
Obligatory https://m.xkcd.com/1205
I am not sure if it was simon or pg who might've quoted this but I remembered a quote about that a 2 magnitude order in speed (quantity) is a huge qualititative change in it of itself.
Not every use case of jq is a person using it interactively in their terminal, believe it or not.
If somebody needs performance, they probably shouldn't be calling out to a separate process for json of all things, no?
(Honestly, who even still writes shell scripts? Have a coding agent write the thing in a real scripting language at least, they aren't phased by the boilerplate of constructing pipelines with python or whatever. I haven't written a shell script in over a year now.)
If you’re writing the script to be used by multiple people, or on multiple systems, or for CI runners, or in containers, etc. then there’s no guarantee of having Python (mostly for the container situation, but still), much less of its version. It’s far too easy to accidentally use a feature or syntax that you took for granted, because who would still be using 3.7 today, anyway? I say this from painful recent experience.
Plus, for any script that’s going to be fetching or posting anything over a network, the LLM will almost certainly want to include requests, so now you either have to deal with dependencies, or make it use urllib.
In contrast, there’s an extremely high likelihood of the environment having a POSIX-compatible interpreter, so as long as you don’t use bash-isms (or zsh-isms, etc.), the script will probably work. For network access, the odds of it having curl are also quite high, moreso (especially in containers) than Python.
If you're distributing the script to other people then the benifit of using python and getting stuff like high quality argument parsing for free is even greater.
If Ms performance is a main concern, you shouldn't use jq. Believe it or not.
For people chewing through 50GB logs or piping JSON through cron jobs all day, a 2x speedup is measurable in wall time and cloud bill, not just terminal-brain nonsense. Most people won't care.
If jq is something you run a few times by hand, a "faster jq" is about as compelling as a faster toaster. A lot of these tools still get traction because speed is an easy pitch, and because some team hit one ugly bottleneck in CI or a data pipeline and decided the old tool was now unacceptable.
I deal with a fair amount of newline-delimited JSON in my day job, where each line in the file is a complete JSON object. I've seen this referred to as "jsonl", and it's not entirely uncommon for logs and other kinds of time-series data dumps. Do any of the popular JSON CLI tools work with this format? I didn't see any mention of it here.
Having used `jq` and `yq` (which followed from the former, in spirit), I have never had to complain about performance of the _latter_ which an order of magnitude (or several) _slower_ than the former. So if there's something faster than `jq`, it's laudable that the author of the faster tool accomplished such a goal, but in the broader context I'd say the performance benefit would be required by a niche slice of the userbase. People who analyse JSON-formatted logs, perhaps? Then again, newline-delimited JSON reigns supreme in that particular kind of scenario, making the point of a faster `jq` moot again.
However, as someone who always loved faster software and being an optimisation nerd, hat's off!
> Having used `jq` and `yq`
If you don't mind me asking, which yq? There's a Go variant and a Python pass-through variant, the latter also including xq and tomlq.
Integrating with server software, the performance is nice to have, as you can have say 100 kRPS requests coming in that need some jq-like logic. For CLI tool, like you said, the performance of any of them is ok, for most of the cases.
jq is probably faster than storage, the network, compression, or something else in your stack and not your bottleneck.
I use jq to grind through gigabytes of GeoJSON files exported from ArcGIS, as an ETL stage. It takes a long time.
Yeah, turns out not everyone uses these tools the way you do. Weird!
I switched to Jaq[0] a while back for the 'correctness' sake rather than performance. But Jaq also claims to be more performant than jq.
[0]: https://github.com/01mf02/jaq
Thank you for the recommendation.
It looks like jaq has already progressed much further in the right direction than jsongrep has just started in the not-quite-as-right direction.
I keep an eye on jaq, but there are some holes in the story. jaq 3.0 is faster than Linux distro builds of jq, but jq built correctly is faster than jaq. As far as I can tell the performance reputation of jq is caused by bad distro packaging.
When initially opening the page it had broken colors in light mode. For anyone else encountering it: switch to dark mode and then back to light mode to fix it.
I suspect the website is vibe-coded, like the tool itself.
I can forgive vibe code... It needs to execute if it works it's fine.
Unedited vibe documentation is unforgivable.
this is a bad faith take. i think the website is really cool and doesn't reek of slop at all. what makes you think differently?
I would not be surprised at all if it's vibe coded. I have seen exactly the same thing myself.
I gave instruction to Claude to add a toggle button to a website where the value needs to be stored in local storage.
It is a very straightforward change. Just follow exactly how it is done for a different boolean setting and you are set. An intern can do that on the first day of their job.
Everything is done properly except that on page load, the stored setting is not read.
Which can be easily discovered if the author, with or without AI tools, has a test or manually goes through the entire workflow just once. I discovered the problem myself and fixed it.
Setting all of that aside -- even if this is not AI coded, at the least it shows the site owner doesn't have the basic care for its visitors to go through this important workflow to check if everything works properly.
Same.
And who cares if it's vibe-coded or not. Since when do we care more on the how than on the what? Are people looking at how a tool was coded before using it, as if it would accelerate confidence?
It’s a heuristic to approach a program a bit warily as the length of the documentation likely outpaces how thoroughly it was designed and tested.
Hey OP here! Sorry about this this is just laziness on my part because I never use light mode so I forget to test haha, will push a fix!
I think this has just been fixed. A bit of dark mode was leaking into light in the css.
I still saw the same bug just now (Firefox on macOS).
It’s still broken for me at this point. White link text on nearly white background. Impossible to read. Safari on my iPad.
Fine in Firefox on Android. Note that the scales of the charts are all different, which makes them hard to compare.
Also, there are lots of charts without comparison so the numbers mean nothing...
Should be fixed now! Let me now :)
White text with light background, yeah.
I had the same problem (brave browser)
Looks fine to me on Edge/Windows.
Modern programmers these days just give a shit about user experience. Better to just load up in reader mode.
Broken on iOS Safari too
I learned a number of data processing cli tools: jq, mlr, htmlq, xsv, yq, etc; to name a few. Not to the level of completing advent of code or anything, but good enough for my day to day usage. It was never ending with the amount of formats I needed to extract data from, and the different syntax's. All that changed when I found nushell though, its replaced all of these tools for me. One syntax for everything, breath of fresh air!
+1. I switched to using Nushell as my daily driver around mid-2023 (0.84.0?) and use it in preference to other interactive tools. I do keep at hand jq, yq, and mlr because I need to exchange stuff with colleagues who don't use Nu.
Same! Nushell replaced almost all of them
Had to spend some efforts to set up completions, also there some small rough edges around commands discoverability, but anyway, much better than the previous oh-my-zsh setup
Ideally, wish it also had a flag to enforce users to write type annotations + compiling scripts as static binaries + a TUI library, and then I'd seriously consider it for writing small apps, but I like and appreciate it in the current state already
Same here, nushell is awesome! It helped me to automate so many more things than I did with any other shell. The syntax is so much more intuitive and coherent, which really helps a lot for someone who always forgot how to write ifs or loops in bash ^^
First of all, congratulations! Nice tool!
Second, some comments on the presentation: the horizontal violin graphs are nice, but all tools have the same colours, and so it's just hard to even spot where jsongrep is. I'd recommend grouping by tool and colour coding it. Besides, jq itself isn't in the graphs at all (but the title of the post made me think it would be!).
Last, xLarge is a 190MiB file. I was surprised by that. It seems too low for xLarge. I daily check 400MiB json documents, and sometimes GiB ones.
Hey thank you! OP here, yes I was struggling to find large enough documents to run the benchmarks on, the range currently on the benchmark data is ~106 B - ~190MB, which I think covers the majority of quick task workloads, but would love to have large documents, if there's an public ones you can thinking of I'd like to know!
The US government tend to offer big public json document [0], such as crime rates [1], or others.
[0]: https://catalog.data.gov/dataset/?res_format=JSON
[1]: https://catalog.data.gov/dataset/crimes-2001-to-present
After reading the title, I was worried that this wasn't written in Rust!
If rust is not in the HN title and fire emojis in the readme, it doesn't come from the Rust region of France.
It's just sparkling memory safe high performance software
From their README [0]:
> Jq is a powerful tool, but its imperative filter syntax can be verbose for common path-matching tasks. jsongrep is declarative: you describe the shape of the paths you want, and the engine finds them.
IMO, this isn't a common use case. The comparison here is essentially like Java vs Python. Jq is perfectly fine for quick peeking. If you actually need better performance, there are always faster ways to parse JSON than using a CLI.
[0]: https://github.com/micahkepe/jsongrep
The data viz of the benchmarks is really rough. I think you’d get a lot of leverage out of rebuilding it and using colors and/or shapes to extract additional dimensions. Nobody wants to scan through raw file paths as labels to try and figure out what the hell the results are
Having the equivalent jq expression in these examples might help to compare expressiveness, and it might help me see if jq could “just” use a DFA when a (sub)query admits one. grep, ripgrep, etc change algorithms based on the query and that makes the speed improvements automatic.
You could just take simdjson, use its ondemand api and then navigate it with .at_path(_with_wildcard) (https://github.com/simdjson/simdjson/blob/master/doc/basics....)
The whole tool would be like a few dozen lines of c++ and most likely be faster than this.
Everything can be written in JavaScript will be written in JavaScript.
Everything can be rewritten in Rust will be written in Rust.
I'm glad you adjusted the CSS while I was typing my comment. I needed to switch to dark mode to be able to read highlighted words.
Nice write up. I will try out your tool.
LOL ... came here to grips about that!
Also "jg" reads very similar to "jq", and initially I thought he was talking about "jq" all along, and I was like: where can I see the "jasongrep" examples? Threw me off for a minute.
Quick question: Isn't the construction of a NFA - DFA a O(2^n) algorithm? If a JSON file has a couple hundred values, its equivalent NFA will have a similar amount, and the DFA will have 2^100 states, so I must be missing something.
theory is one thing but the cpu cache is the real bottleneck here... here is a small visual breakdown of how these arrays look in memory and why pointer chasing is so expensive compared to the actual logic: https://vectree.io/c/json-array-memory-indexing
basically the double jump to find values in the heap is what slows down these tools most
I can see that in practice the bottleneck isn't the automata construction, I'm just curious of how the construction is approached with such a super-exponential conversion algorithm
One problem I have not seen addressed by jq or alterataives, perhaps this one addresses it, is "JSON-like" data. That is, JSON that is not contained in a JSON file
For example, web pages sometimes contain inline "JSON". But as this is not a proper JSON file, jq-style utilties cannot process it
The solution I have used for years is a simple utility written in C using flex^1 (a "filter") that reformats "JSON" on stdin, regardless of whether the input is a proper JSON file or not, into stdout that is line-delimited, human-readable and therefore easy to process with common UNIX utilities
The size of the JSON input does not affect the filter's memory usage. Generally, a large JSON file is processed at the same speed with the same resource usage as a small one
The author here has provided musl static-pie binaries instead of glibc. HN commenters seeking to discredit musl often claim glibc is faster
Personally I choose musl for control not speed
1. jq also uses flex
I am excited for some alternative syntax to jq's. I haven't given much thought to how I'd write a new JSON query syntax if I were writing things from scratch, but I personally never found the jq syntax intuitive. Perhaps I haven't given it enough effort to learn properly.
You don't learn it properly. It's not supposed to be intuitive, it's supposed to be concise at the cost of it being intuitive. Would be like somebody saying typing words in to Google is more intuitive than writing regex.
jq is supposed to fit in to other bash scripts as a one liner. That's it's super power. I know very few people who write regex on the fly either (unless you were using it everyday) they check the documentation and flesh it out when they need it.
Just use Claude to generate the jq expression you need and test it.
Minor suggestion: often I just want to extract one field, whose name I know exactly. I see that `jg` has an option `-F` like this:
$ cat sample.json | jg -F name
I would humbly suggest that a better syntax would be:
$ cat sample.json | jg .name
for a leaf node named "name"; or
$ cat sample.json | jg -F .name.
for any node named "name".
Nowadays I'd just use clickhouse-local / chdb / duckdb to query json files (and pretty much any standard format files)
Since when was jq considered slow?
I've been using jj, which apparently is also faster than jq https://github.com/tidwall/jj
jj was already taken by jujutsu, unfortunately.
I highly recommend anyone to look at jq's VM implementation some time, it's kind of mind-blowing how it works under the hood: https://github.com/jqlang/jq/blob/master/src/execute.c
It does some kind of stack forking which is what allows its funky syntax
The backtracking implementation in jq is really the secret sauce for how it handles those complex filters without getting bogged down
Looks like naive implementation of homemade bytecode interpreter. What's so mind blowing about that? Maybe I missed something.
Surprised to see that there's no official binaries for arm64 darwin. Meaning macOS users will have to run it through the Rosetta 2 translation layer.
OP here: Releases have been updated! Also someone was kind enough to package it in Homebrew already :) https://github.com/micahkepe/jsongrep/pull/22
Just hit this too:
https://news.ycombinator.com/item?id=47542182
The reason I was interested, was adding the new tool to arkade (similar to Brew, but more developer/devops focused - downloads binaries)
The agent found no Arm binaries.. and it seemed like an odd miss for a core tool
https://x.com/alexellisuk/status/2037514629409112346?s=20
I’d install it via cargo anyway and that would build it for arm64.
If the arm64 version was on homebrew (didn’t check if it is but assume not because it’s not mentioned on the page), I’d install it from there rather than from cargo.
I don’t really manually install binaries from GitHub, but it’s nice that the author provides binaries for several platforms for people that do like to install it that way.
You can use cargo-binstall to retrieve Github binary releases if there are any.
Really? That is your response? This is an high quality article from someone who spend a lot of time implementing a cool tool and also sharing the intricate inner workings of it. And your response is, "eh there are no official binaries for my platform". Give them some credit! Be a little more constructive!
His response at least fits the discussion and is relevant to the tool, not generic hollier-than-thou scolding.
To address the concern, anyway, I'm sure it would soon be available in brew as an arm binary.
Since the query compilation needs exponential time, I wonder how large the queries can be before jsongrep becomes slower than all the other tools. In that regard, I think the library could benefit from some functionality for query compilation at compile-time.
Thank you. Very cool. Going to try embedding this into my JSON viewer. One thing I’ve struggled with is that live querying in the UI is constrained by performance.
forgive me my rant, but when I see "just install it with cargo" I immediately lose interest. How many GB do I have to install just to test a little tool? sorry, not gonna do that
I was a bit skeptical at first, but after reading more into jsongrep, it's actually very good. Only did a very quick test just now, and after stumbling over slightly different syntax to jq, am actually quite impressed. Give it a try
What were your syntax stumbling blocks? I must be honest I've used jq enough but can never remember the syntax. It's one of the worst things about jq IMO (not the speed, even though I'm a fan of speedups). There's something ungrokkable about that syntax for me.
Just the basic things, like viewing the complete json (with syntax highlighting) to then determine the filter, that is '.' becomes '**'
Just about to read, but I had to change to dark mode to be able to see the examples, which are bold white on a white background.
Nice.
Some bits of the site are hard to read "takes a query and a JSON input" query is in white and the background of the site is very light which makes it hard to read.
If it's easier to use than jq, they should sell the tool on that.
If the author cares, I can’t read everything on this page. The command snippets have a “BASH” pill in the top left that covers up the command I’m supposed to run. And then there are, I guess topic headings or something that are white-on-white text, so honestly I don’t know what they say or what they are.
OP here: sorry about that, the light mode inconsistencies should be fixed now. Will continue to work on making the site design better as well!
Speed is good! Not a big fan of the syntax though.
Quick comment for the author.
Just added this new tool to arkade, along with the existing jq/yq.
No Arm64 for Darwin.. seriously? (Only x86_64 darwin.. it's a "choice")
No Arm64 for Linux?
For Rust tools it's trivial to add these. Do you think you can do that for the next release?
https://github.com/micahkepe/jsongrep/releases/tag/v0.7.0
OP here- I will add! Thank you for checking out the project!
They've been added!
Is Jq slow?
no
Another alternative is oj, https://github.com/ohler55/ojg. I don't know how the performance compares to jq or any others but it does use JSONPath as the query language. It has a few other options for making nicely formatted JSON and colorizing JSON.
Reminder you can also get DuckDB to slurp the JSON natively and give you a much more expressive query model than anything jq-like.
How does it deal with nested objects? E.g. one of the fields/columns is an array of objects.
Effort would be better investigated making `jq` itself faster.
Many Useless Uses of cat in this documentation. You never need to do `cat file | foo`, you can just do `<file foo`. cat is for concatenating inputs, you never need it for a single input.
As someone who worked with Unix/Linux and command line arguments for 30 years and still "abuse" cat like the documentation, I regularly hear this complaint.
Yes, "cmd <file" is more efficient for the computer but not for the reader in many cases. I read from left to the right and the pipeline might be long or "cmd" might have plenty of arguments (or both). Having "cat file | cmd" immediately gives me the context for what I am working with and corresponds well with "take this file, do this, then that, etc" with it) and makes it easier for me to grok what is happening (the first operation will have some kind of input from stdin). Without that, the context starts with the (first) operation like in the sentence "do this operation, on this file (,then this, etc)". I might not be familiar with it or knowing the arguments it expects.
At least for me, the first variant comes more naturally and is quicker to follow (in most cases), so unless it is performance sensitive that is what I end up with (and cat is insanely fast for most cases).
If left-to-right is your main concern, observe that the post you replied to uses
which is equivalent toThe fastest alternative to jq is to not use JSON.