 # How to write a simple function ?

Hi ! I just began to learn how to use R because of my studies in economy. But I'm facing a "stupid" problem : I don't know how I am supposed to write a function like this : Sooo... I'd like to know which command I have to write to see his graph ? x can be any value... Also, I guess I have to do something to "limit" the graph at the risk of seeing an infinite line ? Or maybe R does it by himself ?

I hope I'm clear enough thank you in advance !

Hi, welcome!

Homework inspired questions are welcome here, but you have to tell us what have you tried so far? what is your specific problem? We are more inclined towards helping you with specific coding problems rather than doing your work for you.

well ; the thing is I put this on general because I'm not really asking about a homework, but about knowing how to create a function's graph on Rstudio it could be this one or another, I just don't know which command is required to write a function...? I tried to understand on wikipedia but It didn't help much Here is the syntax for writing a trivial 'identity' function.

``````identifyfunc <- function(x){
x
}

#testing it
identifyfunc(5)
identifyfunc("hello")
``````

and also the syntax for doing conditional operations on a vector (using dplyr package)

``````library(dplyr)
(x <- 1:50)
case_when(
x %% 35 == 0 ~ "fizz buzz",
x %% 5 == 0 ~ "fizz",
x %% 7 == 0 ~ "buzz",
TRUE ~ as.character(x)
)``````

Much appreciated ! Thank you Defining a function and plotting a function are two related but different topics, since Nir is already addressing the first part here I suggest you open a new topic for the second one providing a REPRoducible EXample (reprex).

About the first part, here you can find a nice introduction to functions in R

steps:

1. define a function `f()`
2. produce a set of paired points with it `(x, y = f(x))`
3. plot them

For the first step you can use the function `case_when()` from
the package `{dplyr}`, and create a standard R function, following
material already suggested, and looking at the internal helpo of R
(i.e., `?case_when`)

For the second one you can take advantage from the `{tibble}` package
(which is loaded by `{dplyr}`, so you do not need to `library()` it
explicitely). That is not mandatory, but a `tibble` has few advantage
against standard `data.frames` you can explore whats and whys
here

For the third last step, you would love `{ggplot2}` packages. That
implements the effective “Grammar of graphics” to draw a statistical/
data science graph. At the very first level, you would start from the
data (a `tibble`); next you would define the link between information
into the data (i.e., column/variable) and information into the plot
(i.e., the so-called aesthetics: x, y, colour, …); and you would
decide how do you what to draw them (i.e. which _geom_etries would
you like to use). That is a very flexible, easy-to-use and
easy-to-learn tools I strongly suggest you to explore.

You can find a great instruments to learn all of this stuff in the
R 4 data science free-online book
(there there is all you need to start effectively to work with data
in R: cap 3: data visualization, cap. 5: data transformation,
cap. 10: tibbles, cap. 19 functions).

Here below, I report to you my purpose to solve you question,
considering all the suggestion you have already obtained by other.
I tried to keep as simple and readable as possible.

``````library(dplyr)
#>
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#>
#>     filter, lag
#> The following objects are masked from 'package:base':
#>
#>     intersect, setdiff, setequal, union
library(ggplot2)

f <- function(x) {
case_when(
(x > 0) & (x < 1/2)  ~ 8/3,
(x >= 1/2) & (x < 1) ~ 4/3,
TRUE                 ~ 0 # TRUE at the end means -> "elsewhere"
)
}

data_to_plot <- tibble(
x = seq(from = -2, to = 3, length.out = 300), # 300 x points
y = f(x)
)

ggplot(data = data_to_plot, mapping = aes(x, y)) +
geom_line()
`````` Created on 2020-09-03 by the reprex package (v0.3.0)

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.