rstudio / cheatsheets Goto Github PK

From @hadley:
Please add an updated devtools/usetthis cheat sheet....the current one is getting a bit old

There is a mistake in the dplyr Cheatsheet (2nd Page), as the mad() R Function corresponds to the "Median of the Absolute Deviations from the Median" (Robust to Outliers).

help(mad)

And not the "Mean Absolute Deviation".

Users might be fooled by Interpretation.

Thank you.

Best,
Juan

The example for table3 is missing the sep argument to separate(table3, rate, into = c("cases", "pop")

should be separate(table3, rate, sep="/", into=c("cases", "pop"))

Same for separate_rows example

I created a Pull Request for Updating the H2O Cheatsheets with fixes from H2O Team.

Juan

the cheat sheet for ggplot "data-visualization-2.1.pdf " doesn't contain ggplot in the title, which when searching for the package's name could be a bit confusing. rename cheat sheet file to 'ggplot2' or 'data-visualization-with-ggplot2-2.1.pdf'?

Now uses .dir argument. https://purrr.tidyverse.org/reference/reduce.html

We're adding thumbnails of the cheat sheets to relevant tidyverse packages, and, in order to avoid increasing package size, @hadley suggested we might host them here.
See tidyverse/stringr#239 (comment)
Would you be amenable to that? If so, I can submit a PR with the images (possibly as a subfolder of /pngs, since that's their file format.

Thanks!

Near the bottom, in the map_dbl explanation, you use n_iris %>% transmute(n=map_int(...)) instead of transmute(n=map_dbl(...)).

Just noticed a small typo in the devtools cheatsheet, 1st page down left corner

devtools::add_build_ignore() is actually
devtools::use_build_ignore()

the cheat sheets are great resource - thanks!

In the ggplot2 Cheat sheet (data-visalization-2.1.pdf), the ggplot2 "logo" in the upper right corner has a very low resolution compared to the rest of the pdf.

In the example code for replace_na() function is the last x2, after the named list, a typo?

It is my understanding that there is no third argument for replace_na(), and the function works fine without it?

From Bill Behrman: It would be good if you could update the faceting section from the older ~ method to the newer vars() method.

Add RStudio Connect to this list of application sharing suggestions/options

Under the Logical and boolean operators to use with filter() section, it reads See ?base::logic, but it should read See ?base::Logic with an upper-case L in Logic.

Switch out devtools::use_*() with usethis::use_*()

Then a few places that could mention another function:

• Setup usethis::use_description()
• Write Code Create a new code file with usethis::use_r()
• Test Create a new test file with usethis::use_test()

Hello,
Data-wrangling cheatsheet is a fantastic every-day used tool ! Congrat !
You could maybe enhance it in adding

1. Within Summarize data, info on summarise_all summarise_at and summarise_if commands that I found usefull to know
2. Within Combine data sets, info on syntax to join by different column names , by=c("x1"="X1") that I often need to use

Footer on the ggplot2/data visualization cheat sheet reads Learn more with browseVignettes(package=c("dplyr", "tibble"))... (pic highlighting here).

Help file says it's the deprecated version and cheatsheet provides no reason to use it over the standard separate_rows().

The Sparklyr Cheatsheet was compiled for sparklyr 0.5 and the current release is 0.9.3. Here are some of the notable issues:

1. In the minimal local example:

• The following function: sdf_predict(model_iris, test_iris)
• Should be updated to: sdf_predict(test_iris, model_iris)
• Console warning message: The signature sdf_predict(model, dataset) is deprecated and will be removed in a future version. Use sdf_predict(dataset, model) or ml_predict(model, dataset) instead.

2. IDE Integration Screenshot:

• What used to be the spark pane is now the connections pane in the RStudio IDE

3. Data Load/Source options with Spark and R:

This workflow that appears at the very top of the cheatsheet implies a troubling workflow in that it shows the data source starting in R. While that is a valid option, it really doesn't communicate that we know the power of connecting to data sources and working with big data and Spark.

I like this graphic that was used in the 0.6 webinar series by Edgar and Javier:

Or this one that shows two different data source paradigms:

4. The graphic on the back page on the bottom right side has been updated to the following (available on spark.rstudio.com):

I have not done a pass over the entire document. These are just a few things I've noticed this week

Is "SAVE FROM SPARK TO FILE SYSTEM" right?

SAVE FROM SPARK TO FILE SYSTEM
Arguments that apply to all functions: x, path
CSV spark_read_csv (header = TRUE, delimiter = ",", quote = """, escape = "\", charset = "UTF-8", null_value = NULL)
JSON spark_read_json (mode = NULL)
PARQUET spark_read_parquet (mode = NULL)

spark_read_[]?

During workshop about geospatial data handling via sf, I gave the sf cheatsheet to participants: it was very useful, thanks for it. It also triggered some ideas for improvement.

We started by reading some .shp files, so we had to use the function st_read() which is not in the cheatsheet. However, loading shape files (.geojson, .shp, ...) in a R session is a quite common and standard operation in geospatial data handling.

Another quite common function to start working with geospatial data is st_as_sf() in order to add geometry to a data.frame by passing the CRS and the columns to use for making a geometry.

What do you think about adding a section in the cheatsheet on how to import geospatial data? Thanks.

I would suggest that the next Cheatsheet in the roadmap is the "reticulate" Cheatsheet:

https://cran.r-project.org/web/packages/reticulate/index.html

Which is already well documented with Vignettes, but I feel it is important too make it more accessible, as R is a language for Interfaces, and it would allow to port many important Python packages to R.

Front page of the ggplot2 cheatsheet, lower right, the heading for the maps section is incorrect... but maybe it's intended as a political statement. :-D

Under the "Summary Functions" heading and the "Counts" subheading, it lists:
sum(!is.na()) - # of non-NA's

However, using the following:
summarise(mtcars, notna = sum(!is.na()))

Produces this error:

#> Error in summarise_impl(.data, dots) : 
#> Evaluation error: 0 arguments passed to 'is.na' which requires 1.

I think the fantastic purrr cheatsheet might need a slight update to rename depth to vec_depth (due to tidyverse/purrr#350). Minimal reprex to demonstrate:

x <- list(list(list(1)))
purrr::depth(x)
#> Error: 'depth' is not an exported object from 'namespace:purrr'
purrr::vec_depth(x)
#> [1] 4

Will there be in the near future a "lubridate" cheatsheet?

It would be greatly appreciated, just as was stringr cheatsheet.

Thank you,
Juan

Some of the quotes on the data transformation cheat sheet were converted to curly quotes (spotted in the join section)

Add a link to each cheatsheet that a user can follow to learn how to use the package (a book or tutorial, once they exist)

Hi All,

would you please consider the correction below?

diff --git a/README.md b/README.md
index 0c176cb..446fef2 100644
--- a/README.md
+++ b/README.md
@@ -1,4 +1,3 @@
-
 ## RStudio Cheat Sheets
 
 <img src="pngs/rstudio-ide.png" width=364 height=288 align="right"/>
@@ -15,7 +14,7 @@ If you wish to contribute to this effort by translating a cht 
sheet, please fe                                                               
 
 ## Tips for making a new cheat sheet
 
-Keep these tips in mind when creating a new cheet sheat:
+Keep these tips in mind when creating a new cheat sheet:
 
 1. RStudio cheat sheets are hosted at https://github.com/rstudio/cheatsheets.ou
 can submit new cheat sheets to the repository with a pull request.            
 

Thank you for making this cheatsheet - very valuable! I just noticed a typo in the assignment of x st_point.

It should be
st_point(x = c(numeric vector), dim = "XYZ"), not
st_point(x, c(numeric vector), dim = "XYZ").

In the stringr cheasheet, str_c function, it says:

str_c(..., sep = "", collapse = NULL) Collapse a vector of strings into a single string.

But the code it's exactly the same as in the paragraph above ('str_c(..., sep = "", collapse = NULL)').

The code should probably be:

str_c(..., sep = "", collapse = "")

Ex. If I want to know how to use tidyr it's not obvious the cheatsheet I need is (the back of) data-import.pdf (and not say data-transformation.pdf)

The links for some cheatsheets at https://www.rstudio.com/resources/cheatsheets/ are not currently working. For example this link to the base-r cheatsheet doesn't currently work:

https://github.com/rstudio/cheatsheets/raw/master/source/pdfs/base-r.pdf

but changing it to point to the old folder does:

https://github.com/rstudio/cheatsheets/raw/master/old/pdfs/base-r.pdf

I think this might be due to this recent commit

Current text reads:

INTERVALS
Divide an interval by a duration to determine its physical length, divide and interval by a period to determine its implied length in clock time.

There is a typo in the second "divide": from "..., divide and ..." to "..., divide an ..."

The description of the arguments is incorrect.

It says:

st_contains(x, y, ... ) identifies if x is within y (i.e. point within a polygon).

However, it should be the reverse. It identifies if y is within x.

Example:
ggplot() + geom_sf(data = test_polygon, alpha = 0.5, fill = "green") + geom_sf(data = my_point, size = 10)

st_contains(x = my_point, y = test_polygon, sparse = FALSE)
[,1]
[1,] FALSE

st_contains(x = test_polygon, y = my_point, sparse = FALSE)
[,1]
[1,] TRUE

The example of table5 joins to columns century and year. But if year is integer 0 or string "0" like in the picture then the result is "200" not "2000". You either need further string operations to convert integer 0 to string 00 or some kind of padding. In any case year should show up as "00".

I think it's not a good example because the "proper" way to do it would be 100*century + year. Better example would be like first and last name.

There is an error in the illustration figure right to column_to_rownames function.
The rowname should be c('t', 'u', 'v') instead of c(1, 2, 3).

I can't find the keynote or ppt or or any source file for the sf.pdf. It has errors and I would like to make changes.

The geom_curve example (b + geom_curve(aes(...)) doesn't work because curvature is assigned to nonexistent z, replacing with a scalar number fixed it, or even defaults.. Older versions had different code that works.

It also sort of implies curvature could vary and @thomasp85 implies in a closed issue that it couldn't be implemented

Instead of 'Combine Tables' the last column on the second page is titled 'Summary Functions'

Trying to open both the keynote and the powerpoint, the font don't seem to work (small things like folder icon). Are the files using font awesome? IF yes, are you using a particular version?

In the regex cheatsheet, there's a typo. ignore.cases should be ignore.case.

I see that the regex cheatsheet is a PowerPoint file. If the easiest way to fix this is for me to download it, make the change, and reupload as a pull request, please let me know.

Thanks, and thanks for creating these great references!

Jake

There are two official templates formats: key for osx users and ppt for windows users.
Is there a chance to add another format that would be more linux/open source friendly?

I think on the dplyr cheatsheet, column "Combine Tables", third but last 'bullet point' it should say by = "col1" instead of by = c("col1", "col2") – even though the latter seems to work to my surprise:
band_members %>% inner_join(band_instruments, by = c("name", "name"))

Also, from the commits I'd guess that the current one has been updated in 2018-03, not 2017-03 (re: footnote)

Hi.

Found a tiny error in the RMarkdown sheet. A newline is missing from the "Customize output with sub-options" section, html_document: should be on the next line.

#79

In the list column cheat sheet, the orange/yellow fonts are near impossible to read when printed in color on white paper. At least on my printer. Perhaps darken or change color scheme?

I have already started doing the h2o R Package cheatsheet. So if anyone is interested in collaboration, I attach what I have done till now:
H2O Cheatsheet v1.04.pptx

I have also uploaded the file to a Github Account I have just created:
https://github.com/jtelleria/H2O-Cheatsheet

In the tidyeval cheasheet, in the "PASS TO ARGUMENT NAMES
OF A QUOTING FUNCTION" section on page 2, the example includes an objected named "name" that doesn't appear anywhere before it is used (in summarise(!!name := mean(!!var)). I think that the line of code above it (var <- rlang::ensym(var)) should probably be name<- rlang::ensym(name), that there should an argument passed to the function called name, and that there should be a var <- rlang::enquo(var) line (like in the "PROGRAM WITH A QUOTING FUNCTION" section example). I may be incorrect, but, in any matter, it seems unclear to me.

In all, I think it should look like this:

named_mean <- function(data, var, name) {
  require(dplyr)
  name <- rlang::ensym(name
  var <- rlang::enquo(var)
  data %>% 
    summarise(!!name := mean(!!var))
}

Currently it's very hard to read

seals$z <- with(seals, sqrt(delta_long^2 + delta_lat^2))l <- ggplot(seals, aes(long, lat))

Also the example is contrived. Why not use a variable that makes sense like z = pop_density instead of whatever delta lat/long calculation is done?

@garrettgman