Mailman 3 July 2013 - Pandas-dev

can we enable dev daily builds ?
by Jeff Reback Aug. 1, 2013

Aug. 1, 2013

2 1

hipchat
by Phillip Cloud July 31, 2013

July 31, 2013

hey guys the chat room is up i'm a little hesitant to post the link on the ML so if you'd like to opt in then send me an email and i'll add you to the list of users. -- Best, Phillip Cloud

1 0

chat
by Phillip Cloud July 31, 2013

July 31, 2013

Hi all, I set up a chat room at http://pandas-dev.hipchat.com i can post the link to join if you guys are interested. i have it for 30 days before i have to start shelling out dough. Wes, do you already have the pandas subdomain? -- Best, Phillip Cloud

3 4

https://github.com/pydata/pandas/pull/4384
by Jeff Reback July 29, 2013

July 29, 2013

This is @jtratners conversion of pandas from 2 to 3 in the codebase itself, incorporating the six module. It alleviates the needs for 2to3 (and makes builds faster). Looks ready to merge to me touches almost every file, so holding up all other PR's until this is in. anyone have final comments...? Jeff

2 1

is there supposed to be a windows 3.3 build on pypi?
by Jeff Reback July 26, 2013

July 26, 2013

I can be reached on my cell (917)971-6387

2 1

0.12
by Jeff Reback July 23, 2013

July 23, 2013

all remaining issues with ujson and reported Debian failures on big endian now closed pending whatever Yaroslavl wants to retest I think good to go for release Jeff

4 8

docs for eval
by Jeff Reback July 22, 2013

July 22, 2013

phillip put up a test of the new eval section in the docs FYI http://phillipcloud.com/doc/build/html/enhancingperf.html#expression-evalua…

1 0

PR's soon after 0.12
by Jeff Reback July 22, 2013

July 22, 2013

All, a couple of big PR's that would like to merge soon after 0.12. Pls take a look and comment. Series refactor: https://github.com/pydata/pandas/pull/3482 comment on this pr for major refactor of Series to be a subclass of NDFrame (and corresponding consolidation of methods et al) Pretty much preserves all APIs (except small changes as noted) and perf. Eval (primarily by @cpcloud): https://github.com/pydata/pandas/pull/4164 provides a nice mechanism to evaluate directly with numexpr as well as expanded syntax e.g. df = DataFrame(randn(100000,2),columns=list('AB')) df['A>0 and B>0'] (this is implemented by df.query(...)) as well as the generic: pd.eval('....') (FYI the allowing of 'and' to be equivalent of '&') is the only real 'API' change here in that its something that neither python nor numpy allow but makes syntax 'nicer' thanks, Jeff

1 0

Docs
by Phillip Cloud July 20, 2013

July 20, 2013

Hi all, I'm wondering if moving the docs over to http://www.rtfd.org might be useful in the long run. It would certainly lift the load on Wes and Chang since they wouldn't have to maintain a separate machine to build docs. Doc builds happen automatically when a push is made (I think you can customize this) and they only show up if the build passes. You can also set up a couple of branches/tags to build from so that users automatically have the option to see stable vs dev docs. -- Best, Phillip Cloud

2 2

Re: [Pandas-dev] Pandas v0.12.1 bugfix release
by Wes McKinney July 19, 2013

July 19, 2013

i'll wait for confirmation and plan to cut the release tomorrow On Thu, Jul 18, 2013 at 2:17 PM, Jeff Reback <jeffreback(a)gmail.com> wrote: > I believe the debian guys has 1 retest on a ujson test, which was merged this morning... > > otherwise I think good to go > > On Jul 18, 2013, at 12:43 AM, Wes McKinney <wesmckinn(a)gmail.com> wrote: > >> On Tue, Jul 16, 2013 at 11:08 AM, Phillip Cloud <cpcloud(a)gmail.com> wrote: >>> I agree with that. Maybe hold off on some of these PRs that are chomping at >>> the bit and make sure things are working relatively swimmingly before >>> opening the gates. >>> >>> >>> -- >>> Best, >>> Phillip Cloud >>> >>> >>> On Tue, Jul 16, 2013 at 2:05 PM, yoval p. <yoval(a)gmx.com> wrote: >>>> >>>> Hi guys, >>>> >>>> As you might remember, there was some confusion post 0.11 on whether >>>> the next release would be 0.12 or 0.11.1, with the result that a >>>> speculative 0.12 became 0.11.1 eliding some borderline changes merged (I >>>> certainly pushed some) which was delayed for so long that it has now >>>> become 0.12 again under it's own weight. >>>> >>>> Seems likely things can work better then that. Might I suggest now >>>> would be a good time to block out a period post 0.12 to address >>>> regressions and bug fixes before moving on to the next evolution? >>>> 2-4 weeks perhaps? >>>> >>>> What do you think? >>>> >>>> yoval >>>> >>>> _______________________________________________ >>>> Pandas-dev mailing list >>>> Pandas-dev(a)python.org >>>> http://mail.python.org/mailman/listinfo/pandas-dev >>> >>> >>> >>> _______________________________________________ >>> Pandas-dev mailing list >>> Pandas-dev(a)python.org >>> http://mail.python.org/mailman/listinfo/pandas-dev >> >> I'll try to get the 0.12 out the door tomorrow afternoon if ya'll give >> me the all clear. Should probably set up a 0.12.x maintenance branch >> where critical bug fixes can be backported for a 0.12.1 if necessary. >> it'd be nice to keep 0.12.x nice and stable for a while. >> _______________________________________________ >> Pandas-dev mailing list >> Pandas-dev(a)python.org >> http://mail.python.org/mailman/listinfo/pandas-dev

2 1