Yoval P. | 15 Jan 13:27 2014

ANN: pandas 0.13.0 released


Pandas v0.13.0 has been released, featuring lots of bug fixes and enhancements
made over 6 months of development in 843 commits from ~70 contributors which
include the core team as well as more first-time contributors then ever and
featuring  our newest team member Joris Van den Bossche. Welcome and Thank you all!

Artifacts have been uploaded to pypi and the full release notes are available at:

Windows binaries are also available from Christoph Gohlke's collection
at: http://www.lfd.uci.edu/~gohlke/pythonlibs/

Now is a good time as any to say a big thank-you to Christoph for providing
such an important resource to the the community.


- New optimized pd.eval() and pd.query() methods fot fats and conise qeury
and expression evaluation by <at> cpcloud and <at> jreback (Experimental).
- a New Float64Index index that extends label based slicing (.ix, .loc) to floats.
- Support for Google BigQuery, contributed by <at> sean-schaefer (Experimental)
- Support for binary serialization to msgpack, <at> jreback (Experimental).
- Tweaks made to clipboard functionality let you seamlessly move data between.
  Pandas and your favorite spreadsheet with ease using read_clipboard and to_clipboard().
- <at> takluyver (ipython team) stopped by and convinced us we should make
  dataframe display more consistent. Regardless of size, DataFrames now consistently
  displays using the data view (with a little added information). As always, you
  can use df.info() to get the info view.
- New df.interpolate method.
- New Convenience methods: str.extract, df.isin().
- Nanosecond support for Offsets.
- Lots of improvements to error messages, let us know about any remaining cryptograms
  you come across.
- <at> alefnula has rifled the bowels of the pandas sandbox to refresh the little-known
  qtpandas, a QT based grid view of DataFrames (GH5013). I ( <at> y-p) had previously made a
  similar jab at a JavaScript equivalent with Exhibitionist and GH2974.
  It seems some sort of GUI df.inspect() method is called for, what do you think?


There are several new sections in the docs <http://pandas.pydata.org/pandas-docs/dev/>
and we're glad to see so much contributions from the community in that area:

- "Comparison with R", Comparing pandas vs. R idiom for common operations.
- "Comparison with SQL", to help SQLers learn the ways of pandas.
- a new "Pandas Ecosystem" that highlights mainly new projects that leverage and
  complement pandas. We hope to see  more new project spring up and engage users and
  contributors and making PyData landscape even more awesome. Get involved!
- a new "Tutorials" section is in construction which will feature all the incredible
  work available out there to help you get your pandas-fu in shape. Please submit a PR
  with your favorites (GH5837).


Finally, the stuff that takes lots of effort and users (hopefully) never see,
there were a a few important changes to internals:
- Led by <at> jtratner, we moved pandas away from 2to3 to a unified code.
- <at> jreback masterly unified series and DataFrame internals under the existing
  NDFrame abstraction. That guy scares me.

Release process

We've been giving much thought to our release process recently and are taking steps
to keep improving on quality and timeliness.

We've consolidated our CI sources into one page (a pain point until now) using
ScatterCI (a side-project of mine). If you have a platform you care about and can
run nightlies or even dedicate a CI box it's now sooo-easy to get your
failing builds on the status page and right in front of the core team's nose
in minutes. If you have a platform you care about and can run a CI box
or just a nightly cron job, inquire within and help us make pandas more stable.
We hope to have SPARC and freebsd builds in the near future.

The new status page lives at: http://scatterci.github.io/pydata/pandas

If things go to plan, 0.14 should be out in 3 months and may or may not include:

- A big upgrade to SQL to/from interop with support for all major DBs by
leveraging SQLAlchemy.
- Template-based displays for dataframes, with conditional formatting and
roll-your-own output generation.
- Reduced memory dataframe construction from known-length iterators.
- Your PRs.

... and much, much more.

Happy new year everyone, help make it a good one!

The Pandas Team


$ λ git log v0.12.0..v0.13.0 --pretty='%aN##%s' | grep -v 'Merge pull' | grep -Po '^[^#]+' | sort | uniq -c | sort -rn
    355 jreback
    114 Phillip Cloud
     83 Jeffrey Tratner
     37 y-p
     21 Jeff Tratner
     21 Andy Hayden
     18 Kieran O'Mahony
     18 Joris Van den Bossche
     12 TomAugspurger
     12 Richard T. Guy
      9 Viktor Kerkez
      9 danielballan
      9 Alex Rothberg
      7 John McNamara
      7 Dan Birken
      6 Thomas Kluyver
      4 Skipper Seabold
      4 PKEuS
      4 John W. O'Brien
      4 Jacob Schaer
      4 Goyo
      4 Andreas Würl
      3 Wes McKinney
      3 rockg
      3 prossahl
      3 Mortada Mehyar
      3 Garrett Drapala
      3 Benedikt Sauer
      2 Yaroslav Halchenko
      2 Valentin Haenel
      2 unutbu
      2 Tiago Requeijo
      2 Kelsey Jordahl
      2 Justin Bozonier
      2 Jan Schulz
      2 Jack Kelly
      2 Ivan Smirnov
      2 DSM
      2 Dieter Vandenbussche
      2 Dale Jung
      2 chapman siu
      2 Ben Alex
      1 zach powers
      1 Zach Dwiel
      1 Wes Turner
      1 westurner
      1 Weston Renoud
      1 Vincent Arel-Bundock
      1 Trent Hauck
      1 Thomas A Caswell
      1 Sten
      1 Roy Hyunjin Han
      1 Roman Pekar
      1 Pierre Haessig
      1 Ondřej Čertík
      1 Olivier Harris
      1 Nick Foti
      1 monicaBee
      1 Mike Kelly
      1 Kyle Meyer
      1 Kyle Kelley
      1 Kyle Hausmann
      1 Kevin Stone
      1 Greg Reda
      1 Gabi Davar
      1 engstrom
      1 daydreamt
      1 David Rasch
      1 d10genes
      1 Christopher Whelan
      1 chappers
      1 Chang She
      1 Caleb Epstein
      1 Brad Buran
      1 Andreas Klostermann
      1 Alex Gaudio
      1 Agustín Herranz

You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
For more options, visit https://groups.google.com/groups/opt_out.