forked from r4ds/bookclub-deepr
-
Notifications
You must be signed in to change notification settings - Fork 0
/
05_vector-indexing.Rmd
341 lines (283 loc) · 7.53 KB
/
05_vector-indexing.Rmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
# Vector indexing
## Basics {-}
### `head()` and `tail()`
+ `head()` and `tail()` return the first or last six elements of an atomic vector
```{r}
x <- 1:10
head(x)
```
```{r}
tail(x)
```
## `'['` - Extract or Replace {-}
+ We can index with the familiar `'['`
+ It's an actual R function!
```{r}
base::'['
```
+ `'['` behaves differently for atomic vectors, lists, data.frames and many other types
```{r}
length(methods('['))
```
+ It has methods for tons of different object classes!
## `'['` - Extract or Replace {-}
```{r}
x <- seq(10, 100, 10)
# First element
x[1]
# Last element
x[length(x)]
# First three elements
x[1:3]
```
## `'['` - Extract or Replace {-}
+ `'['` takes both numeric and logical vectors
+ Therefore, we can index with an expression or the `which()` function
```{r}
x <- runif(10, 1, 10)
y <- runif(10, 1, 10)
# These are the same
all.equal(x[x>y], # indexer returns logical vector
x[which(x>y)]) # indexer returns numeric vector
```
## `'['` - Extract or Replace {-}
+ Note: `which()` by default removes `NA` values, while `'['` does not.
## `'['` - Extract or Replace {-}
+ To *exclude* elements, we can pass negative integers to `'['`
```{r}
x[-1]
```
+ E.g.
```{r}
all.equal(x[-1], x[2:10])
```
## `'['` - Extract or Replace {-}
+ `'['` returns `NA` when we pass `NA` to it
```{r}
x[c(NA, 2, 4)]
```
+ For lists, `'['` returns `NULL` when we pass `NA`
## `'['` - Extract or Replace {-}
+ For named vectors, we can pass characters to `'['`
```{r}
x <- structure(x, names=letters[1:10])
x[c("a", "f", "a", "g", "z")]
```
## `'['` - Extract or Replace {-}
+ `'['` works on lists, and always returns a list
```{r}
y <- list(a = 1:10,
b = 11:20)
y[c(1,2)]
```
## `'[['` and recursiveness {-}
+ `'[['` can be used to select a single element, removing the names attribute.
+ For lists, `'[['` does not return a list but an object of the class of the underlying list element
```{r}
x <- list(a = c(1,2,3),
b = c(4, 5, 6))
data.class(x[1])
```
```{r}
# vs.
data.class(x[[1]])
```
## `'[['` and recursiveness {-}
+ `'[['` operates recursively when given mulitple indexes.
+ Compare:
```{r}
# atomic vector with '['
x <- 1:10
x[c(1, 2)]
```
```{r}
y <- list(a = 1:10,
b = 11:20)
# list with '[['
y[[c(1,2)]] # equivalent to y[[1]][[2]]
```
## `'[<-'` and `'[[<-'` {-}
+ The "replacement versions" of these functions are `'[<-'` and `'[[<-'`
+ These are actually separate functions, that allow us to assign values at a specific index (see Chapter 9)
```{r}
y[[c(1,2)]] <- 999
y
```
## Useful functions {-}
## `match()` and `%in%` {-}
+ `match()` returns the indices of elements in `x` that have matching values in `table`
```{r}
table <- sample(c(1:10, NA), size = 100, replace = TRUE)
x <- c(5, 6, NA)
# returns indices in table
match(x, table, nomatch = 0)
```
+ `%in%` is an intuitive wrapper that returns a logical vector instead
```{r}
# returns logical in x
x %in% table
```
## `findInterval()` {-}
+ `findInterval` returns the interval range in `interval` that contains each element in `x`
```{r}
x <- runif(10)
# a vector of length 4 has 3 intervals
interval <- c(-Inf, .25, .5, Inf)
findInterval(x, interval)
```
## `findInterval() ` {-}
### Exercise 5-21
```{r}
p_winsorised_mean <- function(x, p) {
# calculate quantile cutoffs
x_quantiles <- c(-Inf,
quantile(x, c(p, 1-p), names = FALSE),
Inf)
# map quantiles to x indices
x_intervals <- findInterval(x, x_quantiles)
x_winsorised <- ifelse(x_intervals==2,
x,
ifelse(x == 1, x_quantiles[2], x_quantiles[3]))
return(mean(x_winsorised))
}
x <- c(8, 5, 2, 9, 7, 4, 6, 1, 3)
p <- .25
p_winsorised_mean(x, p)
```
## `split()` and `unsplit()` {-}
+ `split()` can perform grouped aggregations
+ `split()` takes a vector `x`, a factor `f`, and returns a list where each element corresponds to levels of 'f' containing evenly elements of 'x' at corresponding indices.
+ `split()` also tests our knowledge of '[' and '[['!
## `split()` and `unsplit()` {-}
```{r}
x <- c(0, 0.2, 0.25, 0.4, 0.66, 1)
f <- findInterval(x, c(-Inf, 0.25, 0.5, 0.75, Inf))
split(x, f)
```
+ i.e. x[1:2] pairs with f[1:2], x[3:4] pairs with f[3:4]
## `split()` on data.frames {-}
+ Use `Map()` and `split()` to on data.frames to perform grouped aggregations.
```{r}
l <- Map(
function(x) {
c(Min=min(x), Median=median(x), Mean=mean(x), Max=max(x))
},
split(iris[["Sepal.Length"]], iris[["Species"]])
)
print(l[1:3])
```
## vs. Dplyr {-}
```{r}
iris |>
dplyr::group_by(Species) |>
dplyr:: summarise(Min = mean(Sepal.Length),
Median = median(Sepal.Length),
Max = max(Sepal.Length))
```
## vs. Dplyr {-}
+ `split()` may be useful for compatibility with packages that don't have a tidyverse compatible grouping attribute
+ ???
## `order()` and `sort()` {-}
+ `order()` returns a vector that lists *indices* corresponding to the values of 'x'
```{r}
x <- c(10, 20, 30, 50, 40)
order(x)
```
+ `sort()` returns the original vector, in ascending order
```{r}
sort(x)
```
+ `sort(x)` is equivalent to `x[order(x)]`
```{r}
all.equal(sort(x), x[order(x)])
```
## `order()` and `sort()` {-}
+ `order()` can return rankings
```{r}
order(order(x))
# vs.
rank(x, ties = "random")
```
## `duplicated()` {-}
+ `duplicated()` returns a logical vector indicating whether a each element in `x` is repeated in `x`
```{r}
x <- sample(1:10, 10, replace = TRUE)
duplicated(x)
# vs.
match(x, unique(x))
```
## `tabulate()` {-}
+ `tabulate()` returns a count of each element in vector `x`
```{r}
x <- c(2, 4, 6, 2, 2, 2, 3, 6, 6, 3)
tabulate(x)
```
## `tabulate()` {-}
+ (Exercise 5-15)
```{r}
y <- c("a", "b", "a", "c", "a", "d", "e", "e", "g", "g", "c", "c", "g")
result <- structure(tabulate(factor(y)), # y must be a factor or numeric
names = letters[match(unique(y), letters)])
print(result)
```
## Subsetting and Attributes {-}
+ Subsetting functions tend to drop all attributes except for names
+ `'[['` drops *all* attributes
## Applications {-}
+ Replace `NA` values with values from cubic interpolation
+ (Exercise 5-23)
```{r}
euraud <- scan("https://raw.githubusercontent.com/gagolews/teaching-data/master/marek/euraud-20200101-20200630.csv", comment.char = "#")
# create index of NA values
na_index <- is.na(euraud) == TRUE
# assign interpolated values from spline()
# to the NA values using the index
euraud[na_index] <- spline(euraud, n = length(euraud)) |>
_[["x"]] |> # spline() returns a named list
_[na_index]
```
## Applications {-}
+ Reorder a list by it's names attribute
+ (Exercise 5-14)
```{r}
x <- list(a=1, b=2)
y <- list(c=3, a=4)
union_list <- function(x,y){
# combine lists
z <- c(x,y)
# order by names (increasing) then by value (decreasing)
z <- z[order(names(z), unlist(z),
decreasing = c(FALSE, TRUE),
method="radix")]
# remove duplicated names
z[!(duplicated(names(z)) == TRUE)]
}
str(union_list(x,y))
```
## Applications {-}
+ Spearman rank coefficient
```{r}
spearman_rank <- function(x, y) {
n <- ifelse(length(x) == length(y), length(x),
warning("x and y have different lengths"))
r_i = rank(x)
s_i = rank(y)
d_i = r_i - s_i
(6 * sum(d_i^2))/
(n*(n^2-1))
}
x <- rnorm(n = 5)
y <- rnorm(n = 5)
spearman_rank(x,y)
```
## Exercises? {-}
+ Which exercises were the most challenging?
## Meeting Videos {-}
### Cohort 1 {-}
`r knitr::include_url("https://www.youtube.com/embed/URL")`
<details>
<summary> Meeting chat log </summary>
```
LOG
```
</details>