Skip to content

COMPAT: change pytables to use 3.0.0 API (GH7990) #7994

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Aug 12, 2014

Conversation

jreback
Copy link
Contributor

@jreback jreback commented Aug 11, 2014

closes #7990

@jreback jreback added this to the 0.15.0 milestone Aug 11, 2014
@cpcloud
Copy link
Member

cpcloud commented Aug 11, 2014

nice: - is bigger than +

@jreback
Copy link
Contributor Author

jreback commented Aug 11, 2014

that's a first!

@jreback jreback added the CI label Aug 11, 2014
@jreback
Copy link
Contributor Author

jreback commented Aug 11, 2014

@cpcloud @jorisvandenbossche comments?

@@ -13,7 +13,7 @@

try:
import numexpr as ne
_NUMEXPR_INSTALLED = ne.__version__ >= LooseVersion('2.0')
_NUMEXPR_INSTALLED = ne.__version__ >= LooseVersion('2.1')
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

might be nice here to warn if the installed version is < 2.1

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

tho i could see why this being transparent is useful ... just treat it as not installed if the version is incompat...

however as a user of pandas i would like to know that i won't get any numexpr optims when i'm using pandas

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hmm, ok that makes sense I gues.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I mean not a huge deal if you don't like it then don't have to do it.

@cpcloud
Copy link
Member

cpcloud commented Aug 12, 2014

minor comments. otherwise looks good

@jreback
Copy link
Contributor Author

jreback commented Aug 12, 2014

Would this be annoying then? look at the tests

https://travis-ci.org/jreback/pandas/jobs/32325575

@jorisvandenbossche
Copy link
Member

I think it is a useful and informative message, unless you don't bother to use numexpr (and it is eg installed in your distribution) ... then it can become annoying (but that is always like this with warnings no?).
I just looked, I have eg still numexpr 2.0.1 (installed a time ago with anaconda), and I suppose I am not alone (but with conda of course easy to upgrade).

@jorisvandenbossche
Copy link
Member

What is reason for upgrading the required versions? Is it only testing? (that we have to test less versions?) Because both PyTables 3 and Numexpr 2.1 are only 1 to 1.5 year old.

@jreback
Copy link
Contributor Author

jreback commented Aug 12, 2014

@jorisvandenbossche yes < 2.1 is pretty old at this point.

The reason for the changes is motiviated by using conda for ci, as it doesn't support these (pretty old) versions.

And we have to do this at some point before 3.2 PyTables is out (maybe end of year). As the api will start showing a different kind of warning. (it shows DeprecationWarning on all tests atm, but we are suppressing this).

I don't think this is really a big deal. 99% of users dumped 2.4 as soon as 3.0.0 came out because it supports py3 (and had a lot of bug fixes). (and its backward compatible).

@jorisvandenbossche
Copy link
Member

@jreback OK, sounds reasonable

jreback added a commit that referenced this pull request Aug 12, 2014
COMPAT: change pytables to use 3.0.0 API (GH7990)
@jreback jreback merged commit 1eb9955 into pandas-dev:master Aug 12, 2014
@jorisvandenbossche
Copy link
Member

@jreback I got the warning :-)
But I now got: "UserWarning: The installed version of numexpr 2.0.1 is not supported in pandas and will be not be used" -> maybe add what version it should be at least?

@jreback
Copy link
Contributor Author

jreback commented Aug 19, 2014

hmm, was 2.0.0, now its 2.1, ok easy to fix

@jreback
Copy link
Contributor Author

jreback commented Aug 19, 2014

2594bea

@jorisvandenbossche
Copy link
Member

Thanks!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
CI Continuous Integration Compat pandas objects compatability with Numpy or Python functions IO HDF5 read_hdf, HDFStore
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Drop support for PyTables < 3.0.0
3 participants