qplot() ylab behavior changed, now defaults to NULL

[rafalab]

---

With version 3.2.0 qplot, with histogram or density geometries, uses the character NULL as the y label. Previous versions showed nothing.

Update: The difference in behavior seems to be due to the argument .ignore_empty="all" in the call to enquos.

set.seed(2009)
x <- rnorm(100)
qplot(x, geom="histogram")

[yutannihilation]

Thanks, confirmed. As you say, it seems this used to work before #2936, sorry...

packageVersion("ggplot2")
#> [1] '3.1.1'

library(ggplot2)
#> Registered S3 methods overwritten by 'ggplot2':
#>   method         from 
#>   [.quosures     rlang
#>   c.quosures     rlang
#>   print.quosures rlang

set.seed(2009)
x <- rnorm(100)
qplot(x, geom="histogram")
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

^{Created on 2019-08-05 by the reprex package (v0.3.0)}

[yutannihilation]

Now I feel both are wrong; shouldn't the label be count, the same as ggplot() + geom_histogram()...?

[rafalab]

I agree.

qplot(x, geom="histogram") should be the same as
ggplot(aes(x=x), data=data.frame(x=x))+geom_histogrram()

qplot(x, geom="density") should be the same as ggplot(aes(x=x),data=data.frame(x=x))+geom_density()

qplot(sample=x) should be the same as ggplot(aes(sample=x),data=data.frame(x=x))+stat_qq()

and so on.

But the previous behavior of showing nothing is preferable than writing out NULL.

[yutannihilation]

@thomasp85 This is an unexpected change introduced in v3.2.0 (due to my carelessness, sorry...). Do you think reverting this needs to be included in v3.2.1? If so, it would be easy to revert the related changes, though it's not a true fix.

[thomasp85]

I think we should include it... can you make a PR?

[yutannihilation]

Sure, I will send PRs to master and to v3.2.1-rc.

[thomasp85]

just to v3.2.1-rc... It will be merged into master once released

[yutannihilation]

Ah, thanks. I got it.

[lock]

This old issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with reprex) and link to this issue. https://reprex.tidyverse.org/