DOC: Bump fastparquet version by TomAugspurger · Pull Request #24590 · pandas-dev/pandas (original) (raw)
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 andprivacy statement. We’ll occasionally send you account related emails.
Already on GitHub?Sign in to your account
Conversation27 Commits8 Checks0 Files changed
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.Learn more about bidirectional Unicode characters
[ Show hidden characters]({{ revealButtonHref }})
Should we just bump the min version to 0.2.1? It's quite new, but I think reading datetime data is somewhat common :) It'll save us bug reports.
Should we just bump the min version to 0.2.1? It's quite new, but I think reading datetime data is somewhat common :) It'll save us bug reports.
Yeah, maybe that is best ..
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good, let's wait on the doc build to check if it actually worked
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
is this only for docs? may have to update the parquet doc section as well.
Codecov Report
Merging #24590 into master will not change coverage.
The diff coverage isn/a
.
@@ Coverage Diff @@ ## master #24590 +/- ##
Coverage 92.36% 92.36%
Files 166 166
Lines 52497 52497
Hits 48489 48489
Misses 4008 4008
Flag | Coverage Δ | |
---|---|---|
#multiple | 90.79% <ø> (ø) | ⬆️ |
#single | 43.04% <ø> (ø) | ⬆️ |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update db051b9...e9f7841. Read the comment docs.
is this only for docs? may have to update the parquet doc section as well.
For now, this PR is only about the doc build. But if the suggestion of Tom is followed to bump the fastparquet version, we should also update the docs (although, if we decide to not nump the minimum version, we should maybe still mention that for datetime data, the latest version is required)
For now, this PR is only about the doc build. But if the suggestion of Tom is followed to bump the fastparquet version, we should also update the docs (although, if we decide to not nump the minimum version, we should maybe still mention that for datetime data, the latest version is required)
then pls do this here
I've made the changes to bump the min version to 0.2.1 locally, but I think there's an issue with the conda-forge package. Going to fix that before pushing again.
why are you bumping for parquet reading? (we had already bumped for this release I think).
In the discussion above, Joris and I (and I thought you) agreed that reading datetimes is important enough to bump the minimum required first.
On master, we allow using fastparquet <0.2.1, but that will fail at runtime if trying to read a datetimetz column.
The proposal is to require fastparquet>=0.2.1
its fine. we don't usually bump up this fast, but ok.
but would need to change the tests & docs
but fixing the CF package first.
what is this?
No , the problem is not yet fixed (see the doc build log on travis)
Possibly due the conda-forge package problem? (@TomAugspurger the conda list output indicates it installed fastparquet 0.2.1 from conda-forge, but then the error is still there and the pd.sho_versions()
gives fastparquet 0.1.6 ...)
The fastparquet 0.2.1 conda-forge package was accidentally packaging 0.1.6 (but labeling it as 0.2.1), so the error persisted.
The linux builds for a corrected fastparquet are just finishing now. I'll clean up my version bump and push.
Some travis builds are failing (didn't check why), but just checked that the doc build is now working (the error is gone)
The travis build is failing because not all the conda packages are done building. Some are still pulling in the (bad) 0.2.1 fastparquet. Conda install the bad 0.2.1, but then fastparquet.__version__
is 0.1.6 (so we know that code works at least).
I'll let those finish up before pushing again (with the dev env update).
FYI, conda-forge stopped building packages for Python 3.5 on MacOS, so I've removed fastparquet from that env.
Pingviinituutti pushed a commit to Pingviinituutti/pandas that referenced this pull request
Pingviinituutti pushed a commit to Pingviinituutti/pandas that referenced this pull request