(评论)
(comments)

原始链接: https://news.ycombinator.com/item?id=43796898

GS-Calc,一款现代化的电子表格软件,因其处理大型数据集(高达4GB+)和与Python集成能力而在Hacker News上受到关注。用户称赞其在处理数百万行和列时的性能,这比Excel的限制有了显著提升。它提供诸如原地CSV编辑、统计透视数据函数和无限求解变量等功能,并且能够即时创建包含数百万数据点的图表。 评论者强调其在特定任务中取代Excel的潜力,尤其是在处理大型数据集方面,但也承认Excel在兼容性和宏支持方面的优势。评论中也提到了Tad和DuckDB等用于处理大型文件的替代方案。虽然GS-Calc目前仅支持Windows,但计划推出多平台版本。一位用户询问了实时协作功能,开发者提到这可能是未来的扩展功能,但目前没有计划。评论中还提到了另一个基于云的竞争对手Row Zero,它可以处理10亿+行数据。

相关文章
  • Show HN:一款具有Python集成的现代电子表格 2025-04-25
  • (评论) 2024-06-15
  • (评论) 2024-04-30
  • (评论) 2025-03-26
  • (评论) 2024-07-06

  • 原文
    Hacker News new | past | comments | ask | show | jobs | submit login
    Show HN: GS-Calc – A modern spreadsheet with Python integration (citadel5.com)
    98 points by jpiech 1 day ago | hide | past | favorite | 17 comments
    Process large (e.g. 4GB+) data sets in a spreadsheet.

    Load GB/32 million-row files in seconds and use them without any crashes using up to about 500GB RAM.

    Load/edit in-place/split/merge/clean CSV/text files with up to 32 million rows and 1 million columns.

    Use your Python functions as UDF formulas that can return to GS-Calc images and entire CSV files.

    Use a set of statistical pivot data functions.

    Solver functions virtually without limits for the number of variables.

    Create and display all popular chart types with millions of data points instantly.

    Suggestions for improvements are welcome (and often implemented quite quickly).











    It is an interesting tool. I've been struggling with Office Excel's inability to open large files. I always work with csv in python, and if a client must review the data in Excel, I take a random sample to generate a smaller file, then explain to the client that we can't open the whole in Excel. This really doesn't seem like a modern work.

    "a slow, old pc with 8GB RAM"

    By the way, this struck me like a humour of era. Oh god it has 8GB RAM. Cheers! To the good old days.



    I've been using Tad [0] for this purpose, as it streams in the data.

    [0] https://www.tadviewer.com



    Fully agree, I have found that above 300k rows Excel struggles even on a good laptop. Not even mentioning the Python integration into MS Excel that is so unbearably slow that it is much better performing the calculations outside of Excel first.

    I am sold on the website looks and license model!



    These days I use DuckDB to read massive excel files. DuckDB now ships with a nice local UI and it also works beautifully with Datagrip, my preferred database IDE. With SQL, it just becomes a matter of applying old grease to do whatever analysis I want.


    So I'll take a layman's view here since I've only cursory experience of the big data tasks that this software seems to made for. Or maybe the pitch is still different and it went over my head.

    It loads quick, and works with large data. Crucially, you can view and edit visually, not only programmatically.

    Assuming those already working with such data have Excel and Python tools etc., the pitch here is that the $39 license fee saves time or effort. So, is it that the user can spot and correct errors that you couldn't otherwise do with either Excel or with other big data tools? And/or otherwise do the necessary data manipulations?

    I came across the phrase 'eyes like a shithouse rat' recently, to describe the people doing final checks at a printing press. I think there's probably plenty of people out there who would pay $39 for eyes like a shithouse rat.

    Also the website makes me nostalgic :)



    It all sounds very compelling , great work! But I have to ask, what’s the catch? This almost seems like it’s ready to fully replace Excel but I’ve seen many things die in that pursuit. What will Excel users miss by switching?

    I’m don’t do a ton to big data stuff, but sometimes despite Excels stated row and column support- I find it effectively melts down if even 100K/100 of data and forget adding formulas.



    Thanks. This is just a set of features that offer better (sometimes significantly better) performance and make it possible to work with large data sets in some predictable way, that is, knowing what the performance is with 100K formulas, you should be able to estimate how it'll work with 10 millions no matter how much formatting is applied, added hyperlinks, links to external closed workbooks, lists etc. That is probably correct, in the upcoming years there might be no way to compete directly with (the standalone) Excel. Hopefully it just helps to perform or verify various things way faster than in Excel with add-ins (or often at all).


    > What will Excel users miss by switching?

    The ability to use legacy spreadsheets and macros.

    Let's be real, Excel self perpetuates by at once being awful but also the thing everyone used to use thus must still use.

    Lots of spreadsheet apps better than Excel have come and gone over the years...



    MS Office Macros are too easy for untrained users (that think of themselves as power users) to create, and once they grow big enough, it's very hard to port them to other applications. They're the monkey paw of spreadsheets.


    I wonder if anyones gonna eventually just create a wrapper API to just a vm that runs some stock excel open to a spreadsheet...


    VBA and the ability to use Office files with embedded VBA are disabled in many corporations… Some malware in the past used VBA for their attacks and Microsoft never added a proper sandbox.




    are there any plans to add real-time collaboration features, like Google Sheets or Excel Online


    This is a potentially interesting feature - it probably could be relatively easily extended to the shared environment in GS-Calc as there are already separate passwords to protect files, the sheet/view structures and cell ranges and optional sha256 checksums for all the data categories, but there no plans as such. At the moment you can just use two script commands to release the opened file, periodically check whether it's modified, report it and automatically reload the file.


    Looks very cool, but Windows only?


    Thanks. Yes, that was put off a bit too long, but there are plans to come up with the multiplatform version this year.


    https://rowzero.io/ can handle 1 billion+ rows and offers native Python support. Also compatible with Excel and Google Sheets. However it’s a cloud based solution, and the private hosting option is only available to Enterprise users.






    Join us for AI Startup School this June 16-17 in San Francisco!


    Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact



    Search:
    联系我们 contact @ memedata.com