Fixed rendering of ticks in error bars

[drorata]

• closes Error bars not rendered correctly #17467
• tests passed
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

[codecov]

Codecov Report

Merging #17468 into master will decrease coverage by 0.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master   #17468      +/-   ##
==========================================
- Coverage   91.23%   91.21%   -0.02%     
==========================================
  Files         163      163              
  Lines       50102    50102              
==========================================
- Hits        45712    45703       -9     
- Misses       4390     4399       +9

Flag	Coverage Δ
#multiple	89.03% <ø> (ø)	⬆️
#single	40.31% <ø> (-0.06%)	⬇️

Impacted Files	Coverage Δ
pandas/io/gbq.py	25% <0%> (-58.34%)	⬇️
pandas/core/frame.py	97.75% <0%> (-0.1%)	⬇️

---

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 aed9b92...0289098. Read the comment docs.

[drorata]

How did this change changed the coverage?

[drorata]

Any feedback on this PR would be appreciated...

[TomAugspurger]

In #17462 we're discussing changing the style to seaborn or matplotlib 2.0's new defaults. I think both of those also suppress the caps for error bars (by default).

Any reason you want to explicitly add them? Just a style thing?

[drorata]

Well, as is, the code in the example doesn't yield the rendered result. I had (using matplotlib 2.0 and ggplot/default style) to add the cap related option. So this PR is mostly for the sake of correctness of the documentation.

[TomAugspurger]

I think if you were to set `matplotlib.style.use('seaborn')` before making the plot, you'll see the caps. But since we're changing the style used in the docs, I don't think we need the changes here, unless I'm misunderstanding something.

…

On Mon, Sep 11, 2017 at 1:58 PM, Dror Atariah ***@***.***> wrote: Well, as is, the code in the example doesn't yield the rendered result. I had (using matplotlib 2.0 and ggplot/default style) to add the cap related option. So this PR is mostly for the sake of correctness of the documentation. — You are receiving this because your review was requested. Reply to this email directly, view it on GitHub <#17468 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABQHIsLrdrL16Bo32Jmb2qUphX5yd1Icks5shYLGgaJpZM4PQRH-> .

[drorata]

If the change you plan will fix the rendered plot, then you're right. In any case, the current status where the rendered image is not yielded by the sample code is confusing.

[TomAugspurger]

In any case, the current status where the rendered image is not yielded by the sample code is confusing

Yeah it's strange since those aren't static images, they're generated as the code runs. Maybe a version mismatch between your environment and the docs?

Either way I think aesthetically the version without caps looks better. What do you think?

[drorata]

Regarding the environment: could be... I failed to installed the environment as described here but this is a different story I guess.

As for the aesthetic aspect --- I think the caps are crucial for faster and better perception of the error bars.

[pep8speaks]

Hello @drorata! Thanks for updating the PR.

• In the file pandas/io/sql.py, following are the PEP8 issues :

Line 113:80: E501 line too long (80 > 79 characters)

[jorisvandenbossche]

@drorata It seems something went wrong with the latest update, as there is now some unrelated code in the diff.

[drorata]

As per @TomAugspurger comment I guess this PR should be closed and not merged.