Factors with forcats:
CHEATSHEET
The forcats package provides tools for working with factors, which are R's data structure for categorical data.
Factors
R represents categorical data with factors. A factor is an integer vector with a levels attribute that stores a set of mappings between integers and categorical values. When you view a factor, R displays not the integers, but the values associated with them
integer vector"
levels -
stored 1 1 = 0
0 im
1
displayed
□
□
H3 H
ra3H □
□
Create a factor with factorQ
factor(x = character), levels, labels = levels, exclude = NA, ordered = is.ordered(x), nmax = NA) Convert a vector to a factor. Also as_factor. f<-factor(c("a", "c", "b", "a"), levels = c("a", "b", "c")j
Return its levels with levelsQ
levels(x) Return/set the levels of a factor, levels(f); levels(f) <- c("x","y","z")
Use unclass() to see its structure
Inspect Factors
□ l □
H E
H 3
□
□n
01
fct_count(f, sort = FALSE) Countthe number of values with each level. fct_count(f)
Change the order of levels
□ in
H 2=13 03 B
□
1=0
2 = h
□2=n
□ l □
□ i = 0
□2 = h 3=n
□
Q l h H20
I3i a
n2n
Gl
i=0
2-0. 3 = h
□
E
□
■5
fct_relevel(.f,after = OL) Manually reorder factor levels.
fct_relevel(f, c("b", "c" "a"))
fct_infreq(f, ordered = NA) Reorder levels by the frequency in which they appear in the data (highest frequency first). f3 <- factor(c("c", "c", "a")) fct_infreq(f3)
fct_inorder(f, ordered = NA) Reorder levels by order in which they appear in the data.
fct_inorder(f2)
fct_rev(f) Reverse level order. f4 <- factor(c("a","b";,c"))
fct_rev(f4)
fct_shift(f) Shift levels to left or right, wrapping around end.
fct_shift(f4)
Change the value of levels
□ in
H n
H3 H
□
H3H □
□ in
H3 H
□
ra 2 □
H3 Q □
H2 = l
□
□
2 1=2 1 2=1
3
a in
ra 2 n
□
□
□ im
2 - Other
fct_recode(.f,...) Manually change levels. Also fct_relabel which obeys purrr::map syntax to apply a function or expression to each level.
fct_recode(f,v="a",x= "b" z = "c") fct_relabel(f, ~ pasteO("x", .x))
fct_anon(f, prefix ="")) Anonymize levels with random integers. fct_anon(f)
fct_collapse(.f,...) Collapse levels into manually defined groups.
fct_collapse(f,x = c("a", "b"))
fct_lump(f, n, prop, w = NULL, otherjevel = "Other", ties.method = c("min", "average", "first", "last", "random", "max")) Lump together least/most common levels into a single level. Also fct_lump_min. fct_lump(f, n = 1)
ra E
H3H □
H □
H3 H
fct_unique(f) Return the unique values, removing duplicates. fct_unique(f)
Combine Factors
H2=b H2=a
□ i □
H Q
H30
□
O
fct_c(...) Combine factors with different levels. fl <- factor(c("a", "c")) 12 <- factor(c("b", "a"))
fct_c(fl, f2)
fct_unify(fs, levels = lvls_union(fs)) Standardize levels across a list of factors.
fct_unify(list(f2, fl))
Studio
□ iE
ra20 a3 B
13
2=h
3-0
mim
1 = B
2 = B
3 = 0
fct_shuffle(f, n = 1L) Randomly permute order of factor levels.
fct_shuffle(f4)
fct_reorder(.f, .x, .fun=median, .desc = FALSE) Reorder levels by their relationship with another variable.
boxplot(data = iris, Sepal.Width ~ fct_reorder(Species, Sepal.Width))
fct_reorder2(.f, .x, .y, .fun = Iast2,.desc = TRUE) Reorder levels by their final values when plotted with two other variables.
ggplot(data = iris,
aes(Sepal.Width, Sepal.Length, color = fct_reorder2(Species, Sepal.Width, Sepal.Length))) +
geom_smooth()
□ 3
H
□
--
□ in
"2=0
H 3=™h'
□
fct_other(f, keep, drop, otherjevel: "Other") Replace levels with "other."
fct_other(f, keep = c("a", "b"))
Add or drop levels
□ 1=0
EH J
□ 1 □
H2=n
□ 1 b
[JJ2 = |3
0 10    fct_drop(f, only) Drop unused levels. rji 2 = 0 f5<-factor(c("a","b"),c("o""b""x")) f6 <- fct_drop(f5)
0 1 = 0    fct_expand(f,...) Add levels to rjj 2 = j3    a factor. fct_expand(f6, "x")
3 =
□ 1 = 0
fct_explicit_na(f, naJevel="(Missing)"] Assigns a level to NAs to ensure they appear in plots, etc.
fct_explicit_na(factor(c("a", "b", NA)))
RStudio® is a trademark of RStudio, Inc. • CCBYSA RStudio- info@rstudio.com • 844-448-1212 • rstudio.com • Learn more at forcats.tidvverse.ore ■ Diagrams inspired by @LVaudor     forcats 0.3.0« Updated: 2019-02