Wes McKinney已認證帳戶

@wesmckinn

Director of OSS not-for-profit focused on . Created pandas, . PMC. Wrote Python for Data Analysis.

2010年2月加入

推文

你已封鎖 @wesmckinn

你確定要查看這些推文嗎?查看推文並不會將 @wesmckinn 解除封鎖

  1. 釘選的推文
    2月20日

    It's alive: A forum for discussions among developers of different open source data projects. If you are experienced with Discourse and want to help admin / configure / moderate to help this bootstrap let me know over DM.

    還原
  2. 14 小時前

    TIL hoarding/panic buying in German is called "Hamsterkauf" 🐹

    還原
  3. 已轉推
    4月16日

    I wrote some thoughts on why I think , Flight, Parquet and associated projects are a game changer for OLAP. But also for data science, data warehousing, and big data in general. Data is the API and Arrow makes it interoperable:

    還原
  4. 4月15日

    On today’s episode of “Shitty Apocalypse” (Homemade mask c/o my mom)

    還原
  5. 4月15日

    Dear friends: if you got stimulus money, and if it’s practical for you, please don’t send it to Jeff Bezos, spend it in your local community instead

    還原
  6. 已轉推
    4月14日

    Don't miss the interview with Wes McKinney (), one of our keynote speakers at , watch this video and know more about Wes. Watch and share!

    還原
  7. 已轉推
    4月12日

    So it’s only taken two years, but now has a Wikipedia page. Congratulations!

    還原
  8. 4月10日

    A curious aspect of pytest maintainer exodus is that they are 2 out of 3 people being paid by Tidelift A quick look at the nature of the project's PR queue is enough to exhaust even a grizzled maintainer like me.

    還原
  9. 已轉推
    4月10日

    Getting very close to having zero-copy transport of scanner results from the tserver to -enabled clients via unix sockets, fd passing and memfds:

    顯示此對話串
    還原
  10. 4月9日
    還原
  11. 4月9日

    Close quarters steam room conversation with Bruce Springsteen in Helsinki (in 2008)

    還原
  12. 4月8日

    We are still discussing a __dataframe__ interchange interface for Python objects that is not coupled to pandas or another library. The idea is to be able to exchange data without forcing the use of a particular intermediate format More feedback welcome

    還原
  13. 4月7日

    At Ursa Labs we've been investing a lot of time in protecting and Parquet C++ users from security issues and crashes from untrusted or invalid inputs. Read about how we've partnered with 's effort for Continuous Fuzzing

    還原
  14. 已轉推
    4月6日

    Nice writeup on using Drill in , and 🏹 with the assist

    還原
  15. 已轉推
    4月6日

    A post on how monitors performance, and some tips for diagnosing performance regressions when they do occur.

    還原
  16. 4月6日

    There are apparently a number of corporate-paid apartments in my building in Nashville. With current mass layoffs, people are correspondingly having to move out of their apartments. A grim sight

    還原
  17. 4月5日
    還原
  18. 已轉推
    4月4日

    Got Kudu's python client working with the new columnar wire format and conversion to pyarrow. Nice perf results: single core reads 170M cells/sec with arrow vs 3.5M cells/sec with the old construct-a-tuple-per-row code path. New bottleneck is localhost TCP!

    顯示此對話串
    還原
  19. 已轉推
    4月3日

    We're currently translating into 100+ languages and we need native speakers to proofread in their language as soon as possible. If you'd like to help, please reply with the language that you are a fluent/native writer in! h/t

    還原
  20. 4月2日

    I don’t support the other networks, for the record. I just don’t think people should be making excuses for Sean Hannity because they hate Rachel Maddow. Or making excuses for Trump because they hate HRC.

    顯示此對話串
    還原
  21. 4月2日

    We're at a bizarre place in political discourse where right-winger's hatred of left-leaning news networks (CNN, MSNBC, etc.) is so intense that they're willing to forgive Fox hosts etc. for spreading disinformation that is literally getting people killed. Getting people killed.

    顯示此對話串
    還原

看來要一段時間讀取資料。

Twitter 可能已超出負載,或發生暫時性的小問題。請再試一次,或造訪 Twitter 狀態以取得更多資訊。

    你也可能也會喜歡

    ·