-
Notifications
You must be signed in to change notification settings - Fork 38
/
Copy pathbk01_base-functions.html
959 lines (869 loc) · 34.2 KB
/
bk01_base-functions.html
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
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
<!DOCTYPE html>
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta charset="utf-8" />
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<meta name="generator" content="pandoc" />
<meta http-equiv="X-UA-Compatible" content="IE=EDGE" />
<title>Relationship to base and plyr functions</title>
<script src="site_libs/jquery-1.11.3/jquery.min.js"></script>
<meta name="viewport" content="width=device-width, initial-scale=1" />
<link href="site_libs/bootstrap-3.3.5/css/flatly.min.css" rel="stylesheet" />
<script src="site_libs/bootstrap-3.3.5/js/bootstrap.min.js"></script>
<script src="site_libs/bootstrap-3.3.5/shim/html5shiv.min.js"></script>
<script src="site_libs/bootstrap-3.3.5/shim/respond.min.js"></script>
<script src="site_libs/jqueryui-1.11.4/jquery-ui.min.js"></script>
<link href="site_libs/tocify-1.9.1/jquery.tocify.css" rel="stylesheet" />
<script src="site_libs/tocify-1.9.1/jquery.tocify.js"></script>
<script src="site_libs/navigation-1.1/tabsets.js"></script>
<link href="site_libs/highlightjs-9.12.0/default.css" rel="stylesheet" />
<script src="site_libs/highlightjs-9.12.0/highlight.js"></script>
<style type="text/css">code{white-space: pre;}</style>
<style type="text/css">
pre:not([class]) {
background-color: white;
}
</style>
<script type="text/javascript">
if (window.hljs) {
hljs.configure({languages: []});
hljs.initHighlightingOnLoad();
if (document.readyState && document.readyState === "complete") {
window.setTimeout(function() { hljs.initHighlighting(); }, 0);
}
}
</script>
<style type="text/css">
h1 {
font-size: 34px;
}
h1.title {
font-size: 38px;
}
h2 {
font-size: 30px;
}
h3 {
font-size: 24px;
}
h4 {
font-size: 18px;
}
h5 {
font-size: 16px;
}
h6 {
font-size: 12px;
}
.table th:not([align]) {
text-align: left;
}
</style>
<style type = "text/css">
.main-container {
max-width: 940px;
margin-left: auto;
margin-right: auto;
}
code {
color: inherit;
background-color: rgba(0, 0, 0, 0.04);
}
img {
max-width:100%;
}
.tabbed-pane {
padding-top: 12px;
}
.html-widget {
margin-bottom: 20px;
}
button.code-folding-btn:focus {
outline: none;
}
summary {
display: list-item;
}
</style>
<style type="text/css">
/* padding for bootstrap navbar */
body {
padding-top: 60px;
padding-bottom: 40px;
}
/* offset scroll position for anchor links (for fixed navbar) */
.section h1 {
padding-top: 65px;
margin-top: -65px;
}
.section h2 {
padding-top: 65px;
margin-top: -65px;
}
.section h3 {
padding-top: 65px;
margin-top: -65px;
}
.section h4 {
padding-top: 65px;
margin-top: -65px;
}
.section h5 {
padding-top: 65px;
margin-top: -65px;
}
.section h6 {
padding-top: 65px;
margin-top: -65px;
}
.dropdown-submenu {
position: relative;
}
.dropdown-submenu>.dropdown-menu {
top: 0;
left: 100%;
margin-top: -6px;
margin-left: -1px;
border-radius: 0 6px 6px 6px;
}
.dropdown-submenu:hover>.dropdown-menu {
display: block;
}
.dropdown-submenu>a:after {
display: block;
content: " ";
float: right;
width: 0;
height: 0;
border-color: transparent;
border-style: solid;
border-width: 5px 0 5px 5px;
border-left-color: #cccccc;
margin-top: 5px;
margin-right: -10px;
}
.dropdown-submenu:hover>a:after {
border-left-color: #ffffff;
}
.dropdown-submenu.pull-left {
float: none;
}
.dropdown-submenu.pull-left>.dropdown-menu {
left: -100%;
margin-left: 10px;
border-radius: 6px 0 6px 6px;
}
</style>
<script>
// manage active state of menu based on current page
$(document).ready(function () {
// active menu anchor
href = window.location.pathname
href = href.substr(href.lastIndexOf('/') + 1)
if (href === "")
href = "index.html";
var menuAnchor = $('a[href="' + href + '"]');
// mark it active
menuAnchor.parent().addClass('active');
// if it's got a parent navbar menu mark it active as well
menuAnchor.closest('li.dropdown').addClass('active');
});
</script>
<!-- tabsets -->
<style type="text/css">
.tabset-dropdown > .nav-tabs {
display: inline-table;
max-height: 500px;
min-height: 44px;
overflow-y: auto;
background: white;
border: 1px solid #ddd;
border-radius: 4px;
}
.tabset-dropdown > .nav-tabs > li.active:before {
content: "";
font-family: 'Glyphicons Halflings';
display: inline-block;
padding: 10px;
border-right: 1px solid #ddd;
}
.tabset-dropdown > .nav-tabs.nav-tabs-open > li.active:before {
content: "";
border: none;
}
.tabset-dropdown > .nav-tabs.nav-tabs-open:before {
content: "";
font-family: 'Glyphicons Halflings';
display: inline-block;
padding: 10px;
border-right: 1px solid #ddd;
}
.tabset-dropdown > .nav-tabs > li.active {
display: block;
}
.tabset-dropdown > .nav-tabs > li > a,
.tabset-dropdown > .nav-tabs > li > a:focus,
.tabset-dropdown > .nav-tabs > li > a:hover {
border: none;
display: inline-block;
border-radius: 4px;
}
.tabset-dropdown > .nav-tabs.nav-tabs-open > li {
display: block;
float: none;
}
.tabset-dropdown > .nav-tabs > li {
display: none;
}
</style>
<!-- code folding -->
<style type="text/css">
#TOC {
margin: 25px 0px 20px 0px;
}
@media (max-width: 768px) {
#TOC {
position: relative;
width: 100%;
}
}
.toc-content {
padding-left: 30px;
padding-right: 40px;
}
div.main-container {
max-width: 1200px;
}
div.tocify {
width: 20%;
max-width: 260px;
max-height: 85%;
}
@media (min-width: 768px) and (max-width: 991px) {
div.tocify {
width: 25%;
}
}
@media (max-width: 767px) {
div.tocify {
width: 100%;
max-width: none;
}
}
.tocify ul, .tocify li {
line-height: 20px;
}
.tocify-subheader .tocify-item {
font-size: 0.90em;
}
.tocify .list-group-item {
border-radius: 0px;
}
</style>
</head>
<body>
<div class="container-fluid main-container">
<!-- setup 3col/9col grid for toc_float and main content -->
<div class="row-fluid">
<div class="col-xs-12 col-sm-4 col-md-3">
<div id="TOC" class="tocify">
</div>
</div>
<div class="toc-content col-xs-12 col-sm-8 col-md-9">
<div class="navbar navbar-default navbar-fixed-top" role="navigation">
<div class="container">
<div class="navbar-header">
<button type="button" class="navbar-toggle collapsed" data-toggle="collapse" data-target="#navbar">
<span class="icon-bar"></span>
<span class="icon-bar"></span>
<span class="icon-bar"></span>
</button>
<a class="navbar-brand" href="index.html">purrr tutorial</a>
</div>
<div id="navbar" class="navbar-collapse collapse">
<ul class="nav navbar-nav">
<li>
<a href="index.html">Lessons and examples</a>
</li>
<li>
<a href="more-resources.html">More resources</a>
</li>
<li>
<a href="talks.html">Talks</a>
</li>
</ul>
<ul class="nav navbar-nav navbar-right">
<li>
<a href="about.html">About</a>
</li>
</ul>
</div><!--/.nav-collapse -->
</div><!--/.container -->
</div><!--/.navbar -->
<div class="fluid-row" id="header">
<h1 class="title toc-ignore">Relationship to base and plyr functions</h1>
</div>
<div id="why-not-base" class="section level2">
<h2>Why not base?</h2>
<p>You need a way to iterate in R in a data-structure-informed way. What does that mean?</p>
<ul>
<li>Iterate over elements of a list</li>
<li>Iterate over rows or columns of a 2-dimensional object<br />
</li>
<li>Iterate over sub data frames induced by one or more factors</li>
<li>Iterate over tuples formed from the i-th element of several vectors of equal length</li>
</ul>
<p><strong>All of this is absolutely possible with base R</strong>, using <code>for()</code> loops or the “apply” functions, such as <code>apply()</code>, <code>[slvmt]apply()</code>, and <code>by()</code>. I know, because I did so religiously for over 15 years. It works fine!</p>
<p>Why might you do otherwise? As an instructor, I encounter lots of useRs who sheepishly say they know they should be using the apply functions, but they don’t. Because they’ve never quite figured them out or been able to form the habit.</p>
<p>I have a theory about why. The user interface of the “apply”" functions is not as consistent as it could be, which slows down learning. The return objects frequently require further checking and massage to use downstream, which diminishes the payoff. In particular, there’s a tendency to return a vector (atomic or otherwise) or array, instead of data frame, with the original factor levels appearing in a names attribute.</p>
<p>Regardless of your preference, hopefully the side-by-side examples below are a helpful tour of different ways to iterate. Even if you prefer purrr/tidyverse in some settings, you might want base methods when you don’t want any dependencies.</p>
</div>
<div id="why-purrr" class="section level2">
<h2>Why purrr?</h2>
<p>purrr addresses some of the friction identified in the base functions for “split-apply-combine”:</p>
<ul>
<li>The <code>map()</code> family of functions is highly internally consistent, making it easier to transfer expertise from one function to another.</li>
<li>Greater encouragement for type-safe simplification to atomic vector or data frame, producing output that is more ready for the next step.</li>
<li>Concise syntax for defining anonymous functions.</li>
</ul>
<p>We load purrr now. Below we also use example lists from <a href="https://github.com/jennybc/repurrrsive">repurrrsive</a> which are <a href="ls00_inspect-explore.html">introduced and explored elsewhere</a>.</p>
<pre class="r"><code>library(purrr)
library(repurrrsive)</code></pre>
</div>
<div id="why-not-plyr" class="section level2">
<h2>Why not plyr?</h2>
<p>If you have never heard of the <a href="https://CRAN.R-project.org/package=plyr">plyr package</a>, you can skip this section! Personally, I’m a bit sad that it’s time to move on from plyr. But it is time.</p>
<p>Why? plyr is no longer under active development. The innovation is happening elsewhere, in purrr and the other packages in the tidyverse. Also, even I must admit that plyr can be rather slow. I wouldn’t recommend writing new code that requires plyr.</p>
<p>In terms of understanding or translating plyr code, here’s the minimum you need to know: The key feature of the main 3 * 4 = 12 functions in plyr is their explicit, memorable form.</p>
<pre class="r"><code>XYply()</code></pre>
<p>where <code>X</code> and <code>Y</code> specify the type of input and output, respectively. <code>X</code> and <code>Y</code> take values from here:</p>
<ul>
<li><code>a</code> = “array”</li>
<li><code>d</code> = “data frame”</li>
<li><code>l</code> = “list”</li>
<li><code>_</code> = “nothing” (not eligible as <code>X</code> re: input, for obvious reasons)</li>
</ul>
<p>The most useful function proved to be <code>ddply()</code>: data frame in, data frame out. This inspired the creation of the <a href="https://dplyr.tidyverse.org">dplyr package</a>. Most tasks suitable for <code>a*ply()</code> and, especially <code>l*ply()</code> are now handled by purrr’s <code>map()</code> family. The functions called for their side effects, as opposed to return value, are those named <code>*_ply()</code>; this is handled by purrr’s <code>walk()</code> functions. Marching through a data frame row by row is now a job for purrr’s <code>pmap()</code>.</p>
</div>
<div id="lapply-vs.-purrrmap" class="section level2">
<h2><code>lapply()</code> vs. <code>purrr::map()</code></h2>
<p>These are the core mapping functions of base and purrr, respectively. They are “list in, list out”. The main (only?) difference is access to purrr’s shortcuts for indexing by name or position and for creating anonymous functions.</p>
<div class="row">
<div class="col-md-6">
<p><strong>base</strong></p>
<pre class="r"><code>lapply(got_chars[1:3],
function(x) x[["name"]])
#> [[1]]
#> [1] "Theon Greyjoy"
#>
#> [[2]]
#> [1] "Tyrion Lannister"
#>
#> [[3]]
#> [1] "Victarion Greyjoy"</code></pre>
</div>
<div class="col-md-6">
<p><strong>purrr</strong></p>
<pre class="r"><code>map(got_chars[1:3], "name")
#> [[1]]
#> [1] "Theon Greyjoy"
#>
#> [[2]]
#> [1] "Tyrion Lannister"
#>
#> [[3]]
#> [1] "Victarion Greyjoy"</code></pre>
</div>
</div>
<p>This is also <code>plyr::llply()</code>, although it offers some other arguments, such as <code>.progress</code> and <code>.parallel</code>.</p>
<pre class="r"><code>plyr::llply(got_chars[1:3], function(x) x[["name"]])</code></pre>
</div>
<div id="sapply-vs.-__" class="section level2">
<h2><code>sapply()</code> vs. ¯\_(ツ)_/¯</h2>
<p><code>sapply()</code> is a base function that attempts to apply a reasonable simplification to the output of <code>lapply()</code>. It’s handy for interactive use, but due to the unpredictability of it return value, it’s unwise to use it in programming. There is no equivalent in purrr or plyr.</p>
<p>Problem demonstration: The characters in Game of Thrones can have aliases. Some have several, some have one, some have none. Depending on which characters I work with, the same <code>sapply()</code> call can return an object of entirely different class, i.e. a list or a character vector. This can be a source of mysterious bugs in code meant to run non-interactively.</p>
<pre class="r"><code>aliases1 <- sapply(got_chars[20:22], function(x) x[["aliases"]])
str(aliases1)
#> List of 3
#> $ : chr [1:3] "Ned" "The Ned" "The Quiet Wolf"
#> $ : chr [1:4] "The Kingslayer" "The Lion of Lannister" "The Young Lion" "Cripple"
#> $ : chr "Griffthe Mad King's Hand"
aliases2 <- sapply(got_chars[c(3, 22, 27)], function(x) x[["aliases"]])
str(aliases2)
#> chr [1:3] "The Iron Captain" "Griffthe Mad King's Hand" ...</code></pre>
<p>With purrr, you would use <code>map()</code> to get a list back or <code>map_chr()</code> to get atomic character vector. If you use <code>map_chr()</code> when you should not, you’ll get an informative error right away (shown below) and can adjust your approach accordingly.</p>
<pre class="r"><code>map_chr(got_chars[2:4], "aliases")
#> Result 1 must be a single string, not a character vector of length 11</code></pre>
</div>
<div id="vapply-vs.-map_" class="section level2">
<h2><code>vapply()</code> vs. <code>map_*()</code></h2>
<p>Base <code>vapply()</code> requires you to specify a template for the return value and is described as a safer alternative to <code>sapply()</code>. The closest purrr functions are the type-specific mapping functions: <code>map_lgl()</code>, <code>map_int()</code>, <code>map_dbl()</code>, and <code>map_chr()</code> that are “list in, atomic vector out”. Here’s comparable use of <code>vapply()</code> and <code>map_chr()</code> to get some of the Game of Thrones characters’ names.</p>
<div class="row">
<div class="col-md-6">
<p><strong>base</strong></p>
<pre class="r"><code>vapply(got_chars[1:3],
function(x) x[["name"]],
character(1))
#> [1] "Theon Greyjoy" "Tyrion Lannister" "Victarion Greyjoy"</code></pre>
</div>
<div class="col-md-6">
<p><strong>purrr</strong></p>
<pre class="r"><code>map_chr(got_chars[1:3], "name")
#> [1] "Theon Greyjoy" "Tyrion Lannister" "Victarion Greyjoy"</code></pre>
</div>
</div>
<div id="vapply-always-simplifies" class="section level3">
<h3><code>vapply()</code> always simplifies</h3>
<p>What’s not to love with <code>vapply()</code> then? It suffers from the <code>drop = TRUE</code> vs <code>FALSE</code> problem we have when requesting a single row or column from a 2-dimensional object. Except <code>vapply()</code> has no <code>drop</code> argument to control this behavior. It’s an example of the base functions being more difficult to program around. The template allows you to specify the form of each individual result, but there is no way to specify the form – such as the dimension – of the <em>overall</em> result.</p>
<p>I adapt this example from my real life, where I have <code>vapply()</code> inside a function and <code>n</code> is an argument to that function, i.e. it varies. Here I simply define <code>n</code> in the global environment prior to the <code>vapply()</code> call. Note how <code>vapply()</code> returns a 2 dimensional object in the first case and atomic vector in the second. As it says in the docs: “Simplification is always done in <code>vapply</code>.” Believe it.</p>
<pre class="r"><code>f <- function(x, n) rep(x, n)
n <- 3
vapply(c("a", "b"), f, character(n), n = n)
#> a b
#> [1,] "a" "b"
#> [2,] "a" "b"
#> [3,] "a" "b"
n <- 1
vapply(c("a", "b"), f, character(n), n = n)
#> a b
#> "a" "b"</code></pre>
</div>
</div>
<div id="_-vs.-map_dfr" class="section level2">
<h2>¯\_(ツ)_/¯ vs. <code>map_dfr()</code></h2>
<p>The <code>purrr::map_dfr()</code> function is “list in, data frame out” and there is no true base equivalent. Given the centrality of data frames for analysis, it is handy to have a function to produce them, without resorting to <code>do.call()</code>.</p>
<div class="row">
<div class="col-md-6">
<p><strong>base</strong></p>
<pre class="r"><code>l <- lapply(got_chars[23:25],
`[`, c("name", "playedBy"))
mat <- do.call(rbind, l)
(df <- as.data.frame(mat, stringsAsFactors = FALSE))
#> name playedBy
#> 1 Jon Snow Kit Harington
#> 2 Aeron Greyjoy Michael Feast
#> 3 Kevan Lannister Ian Gelder</code></pre>
</div>
<div class="col-md-6">
<p><strong>purrr</strong></p>
<pre class="r"><code>map_dfr(got_chars[23:25],
`[`, c("name", "playedBy"))
#> # A tibble: 3 x 2
#> name playedBy
#> <chr> <chr>
#> 1 Jon Snow Kit Harington
#> 2 Aeron Greyjoy Michael Feast
#> 3 Kevan Lannister Ian Gelder</code></pre>
</div>
</div>
<p>The base workflow above gets trickier if you’re extracting elements of disparate type. At that point, it may make more sense to use <code>vapply()</code> repeatedly. For comparability, we’ll show similar using purrr’s type-specific mapping, which is also safer than relying on automatic type conversion from <code>map_dfr()</code>.</p>
<div class="row">
<div class="col-md-6">
<p><strong>base</strong></p>
<pre class="r"><code>data.frame(
name = vapply(got_chars[23:25], `[[`,
character(1), "name"),
id = vapply(got_chars[23:25], `[[`,
integer(1), "id"),
stringsAsFactors = FALSE
)
#> name id
#> 1 Jon Snow 583
#> 2 Aeron Greyjoy 60
#> 3 Kevan Lannister 605</code></pre>
</div>
<div class="col-md-6">
<p><strong>purrr</strong></p>
<pre class="r"><code>tibble::tibble(
name = map_chr(got_chars[23:25], "name"),
id = map_int(got_chars[23:25], "id")
)
#> # A tibble: 3 x 2
#> name id
#> <chr> <int>
#> 1 Jon Snow 583
#> 2 Aeron Greyjoy 60
#> 3 Kevan Lannister 605</code></pre>
</div>
</div>
</div>
<div id="mapply-vs.-map2-pmap" class="section level2">
<h2><code>mapply()</code> vs. <code>map2()</code>, <code>pmap()</code></h2>
<p>When you need to iterate over 2 or more vectors/lists in parallel, the base option is <code>mapply()</code>. Unlike the other apply functions, the first argument is <code>FUN</code>, the function to apply, and the multiple vector inputs are provided “loose” via <code>...</code>.</p>
<p>For exactly two vector inputs, purrr has <code>map2()</code>, with all the usual type-specific variants. For an arbitrary number of vector inputs, use purrr <code>pmap()</code> or type-specific variants, with the inputs packaged in a list. A very handy special case is when the input is a data frame, in which case <code>pmap_*()</code> applies <code>.f</code> to each row.</p>
<div class="row">
<div class="col-md-6">
<p><strong>base</strong></p>
<pre class="r"><code>nms <- vapply(got_chars[16:18],
`[[`, character(1), "name")
birth <- vapply(got_chars[16:18],
`[[`, character(1), "born")
mapply(function(x, y) paste(x, "was born", y),
nms, birth)
#> Brandon Stark
#> "Brandon Stark was born In 290 AC, at Winterfell"
#> Brienne of Tarth
#> "Brienne of Tarth was born In 280 AC"
#> Catelyn Stark
#> "Catelyn Stark was born In 264 AC, at Riverrun"</code></pre>
</div>
<div class="col-md-6">
<p><strong>purrr</strong></p>
<pre class="r"><code>nms <- got_chars[16:18] %>%
map_chr("name")
birth <- got_chars[16:18] %>%
map_chr("born")
map2_chr(nms, birth, ~ paste(.x, "was born", .y))
#> [1] "Brandon Stark was born In 290 AC, at Winterfell"
#> [2] "Brienne of Tarth was born In 280 AC"
#> [3] "Catelyn Stark was born In 264 AC, at Riverrun"
## and again, but with pmap()
df <- tibble::tibble(
nms,
connector = "was born",
birth
)
pmap_chr(df, paste)
#> [1] "Brandon Stark was born In 290 AC, at Winterfell"
#> [2] "Brienne of Tarth was born In 280 AC"
#> [3] "Catelyn Stark was born In 264 AC, at Riverrun"</code></pre>
</div>
</div>
</div>
<div id="aggregate-vs.-dplyrsummarize" class="section level2">
<h2><code>aggregate()</code> vs. <code>dplyr::summarize()</code></h2>
<p>Consider a data frame, as opposed to a nested list. How do you split it into pieces, according to one or more factors, apply a function to the pieces, and combine the results?</p>
<p>Create a tiny excerpt of the Gapminder dataset that contains a bit of data for Canada and Germany. Load dplyr, now that we are more in the data frame world.</p>
<pre class="r"><code>library(dplyr)
library(gapminder)
(mini_gap <- gapminder %>%
filter(country %in% c("Canada", "Germany"), year > 2000) %>%
droplevels())
#> # A tibble: 4 x 6
#> country continent year lifeExp pop gdpPercap
#> <fct> <fct> <int> <dbl> <int> <dbl>
#> 1 Canada Americas 2002 79.8 31902268 33329.
#> 2 Canada Americas 2007 80.7 33390141 36319.
#> 3 Germany Europe 2002 78.7 82350671 30036.
#> 4 Germany Europe 2007 79.4 82400996 32170.</code></pre>
<p>Simple summary of a single variable for each country. Specifically, take the mean of life expectancy. In this case, the formula method of base <code>aggregate()</code> is quite nice, i.e. it returns a convenient data frame.</p>
<div class="row">
<div class="col-md-6">
<p><strong>base</strong></p>
<pre class="r"><code>aggregate(lifeExp ~ country, mini_gap, mean)
#> country lifeExp
#> 1 Canada 80.2115
#> 2 Germany 79.0380</code></pre>
</div>
<div class="col-md-6">
<p><strong>tidyverse</strong></p>
<pre class="r"><code>mini_gap %>%
group_by(country) %>%
summarize(lifeExp = mean(lifeExp))
#> # A tibble: 2 x 2
#> country lifeExp
#> <fct> <dbl>
#> 1 Canada 80.2
#> 2 Germany 79.0</code></pre>
</div>
</div>
<p>Simple summaries of two variables for each country. We take the mean of life expectancy and of GDP per capita.</p>
<div class="row">
<div class="col-md-6">
<p><strong>base</strong></p>
<pre class="r"><code>## formula method
aggregate(cbind(lifeExp, gdpPercap) ~ country, mini_gap, mean)
#> country lifeExp gdpPercap
#> 1 Canada 80.2115 34824.10
#> 2 Germany 79.0380 31103.09
## data.frame method
aggregate(mini_gap[c("lifeExp", "gdpPercap")], list(mini_gap$country), mean)
#> Group.1 lifeExp gdpPercap
#> 1 Canada 80.2115 34824.10
#> 2 Germany 79.0380 31103.09
## tapply() more general but output less useful here (data frame?)
## returns named vector
tapply(mini_gap$lifeExp, mini_gap$country, mean)
#> Canada Germany
#> 80.2115 79.0380
## returns list
tapply(mini_gap$lifeExp, mini_gap$country, mean, simplify = FALSE)
#> $Canada
#> [1] 80.2115
#>
#> $Germany
#> [1] 79.038</code></pre>
</div>
<div class="col-md-6">
<p><strong>tidyverse</strong></p>
<pre class="r"><code>mini_gap %>%
group_by(country) %>%
summarize_at(vars(lifeExp, gdpPercap), mean)
#> # A tibble: 2 x 3
#> country lifeExp gdpPercap
#> <fct> <dbl> <dbl>
#> 1 Canada 80.2 34824.
#> 2 Germany 79.0 31103.</code></pre>
</div>
</div>
<p>Bivariate summary of two variables for each country. We compute the correlation of life expectancy and year, for the full gapminder dataset now. On the base side, we can no longer use <code>aggregate()</code> or <code>tapply()</code> and need to graduate to <code>by()</code>.</p>
<div class="row">
<div class="col-md-6">
<p><strong>base</strong></p>
<pre class="r"><code>## by() with simplification (the default)
by_obj <- by(gapminder, gapminder$country, function(df) cor(df$lifeExp, df$year))
head(by_obj)
#> gapminder$country
#> Afghanistan Albania Algeria Angola Argentina Australia
#> 0.9735051 0.9542420 0.9925307 0.9422392 0.9977816 0.9897716
## by() without simplification
by_obj <- by(gapminder, gapminder$country, function(df) cor(df$lifeExp, df$year),
simplify = FALSE)
head(by_obj)
#> $Afghanistan
#> [1] 0.9735051
#>
#> $Albania
#> [1] 0.954242
#>
#> $Algeria
#> [1] 0.9925307
#>
#> $Angola
#> [1] 0.9422392
#>
#> $Argentina
#> [1] 0.9977816
#>
#> $Australia
#> [1] 0.9897716</code></pre>
</div>
<div class="col-md-6">
<p><strong>tidyverse</strong></p>
<pre class="r"><code>gapminder %>%
group_by(country) %>%
summarize(cor = cor(lifeExp, year))
#> # A tibble: 142 x 2
#> country cor
#> <fct> <dbl>
#> 1 Afghanistan 0.974
#> 2 Albania 0.954
#> 3 Algeria 0.993
#> 4 Angola 0.942
#> 5 Argentina 0.998
#> 6 Australia 0.990
#> 7 Austria 0.996
#> 8 Bahrain 0.983
#> 9 Bangladesh 0.995
#> 10 Belgium 0.997
#> # … with 132 more rows</code></pre>
</div>
</div>
<p>This is a good place to pause and glance back over the base vs tidyverse calls and results in this section. It shows that you can absolutely do these tasks either way, but the consistency of the code and return values is higher in the tidyverse.</p>
</div>
<div id="by-vs.-tidyrnest" class="section level2">
<h2><code>by()</code> vs. <code>tidyr::nest()</code></h2>
<p>Fit a linear model of life expectancy against year. On the tidyverse side, we now create a nested data frame, with one meta-row per country. Therefore we load tidyr to get <code>nest()</code>. The data needed for each country’s linear model is stored as a list-column of country-specific data frame.</p>
<div class="row">
<div class="col-md-6">
<p><strong>base</strong></p>
<pre class="r"><code>by_obj <- by(gapminder,
gapminder$country,
function(df) lm(lifeExp ~ year, data = df))
str(by_obj[1:2], max.level = 1)
#> List of 2
#> $ Afghanistan:List of 12
#> ..- attr(*, "class")= chr "lm"
#> $ Albania :List of 12
#> ..- attr(*, "class")= chr "lm"
#> - attr(*, "dim")= int 2
#> - attr(*, "dimnames")=List of 1
by_obj[[1]]
#>
#> Call:
#> lm(formula = lifeExp ~ year, data = df)
#>
#> Coefficients:
#> (Intercept) year
#> -507.5343 0.2753</code></pre>
</div>
<div class="col-md-6">
<p><strong>tidyverse</strong></p>
<pre class="r"><code>library(tidyr)
nested_df <- gapminder %>%
group_by(country, continent) %>%
nest() %>%
mutate(fit = map(data, ~ lm(lifeExp ~ year, data = .x)))
str(nested_df$fit[1:2], max.level = 1)
#> List of 2
#> $ :List of 12
#> ..- attr(*, "class")= chr "lm"
#> $ :List of 12
#> ..- attr(*, "class")= chr "lm"
nested_df$fit[[1]]
#>
#> Call:
#> lm(formula = lifeExp ~ year, data = .x)
#>
#> Coefficients:
#> (Intercept) year
#> -507.5343 0.2753</code></pre>
</div>
</div>
<p>What if you want to inspect the fits for Oceania? On the base side, you’ve got some work to do because the country information lives only in the names and the continent information is not directly linked to the model fits at all. On the tidyverse side, where the fits live in a data frame that carries country and continent info, we can use our usual techniques for filtering rows based on the data.</p>
<div class="row">
<div class="col-md-6">
<p><strong>base</strong></p>
<pre class="r"><code>o_countries <- as.character(unique(gapminder$country[gapminder$continent == "Oceania"]))
by_obj[names(by_obj) %in% o_countries]
#> $Australia
#>
#> Call:
#> lm(formula = lifeExp ~ year, data = df)
#>
#> Coefficients:
#> (Intercept) year
#> -376.1163 0.2277
#>
#>
#> $`New Zealand`
#>
#> Call:
#> lm(formula = lifeExp ~ year, data = df)
#>
#> Coefficients:
#> (Intercept) year
#> -307.6996 0.1928</code></pre>
</div>
<div class="col-md-6">
<p><strong>tidyverse</strong></p>
<pre class="r"><code>nested_df %>%
filter(continent == "Oceania") %>%
.$fit
#> [[1]]
#>
#> Call:
#> lm(formula = lifeExp ~ year, data = .x)
#>
#> Coefficients:
#> (Intercept) year
#> -376.1163 0.2277
#>
#>
#> [[2]]
#>
#> Call:
#> lm(formula = lifeExp ~ year, data = .x)
#>
#> Coefficients:
#> (Intercept) year
#> -307.6996 0.1928</code></pre>
</div>
</div>
<p>Let’s form a data frame with one row per country and variables for country, continent, estimated intercept, and estimated slope. I also want to guarantee that the country and continent factors have the same levels as they originally had in gapminder.</p>
<div class="row">
<div class="col-md-6">
<p><strong>base</strong></p>
<pre class="r"><code>coefs <- lapply(by_obj, coef)
coefs <- do.call(rbind, coefs)
coefs <- data.frame(
country = I(rownames(coefs)),
coefs
)
coefs$continent <- gapminder$continent[match(coefs$country, gapminder$country)]
coefs$continent <- factor(coefs$continent, levels = levels(gapminder$continent))
coefs$country <- factor(coefs$country, levels = levels(gapminder$country))
head(coefs)
#> country X.Intercept. year continent
#> Afghanistan Afghanistan -507.5343 0.2753287 Asia
#> Albania Albania -594.0725 0.3346832 Europe
#> Algeria Algeria -1067.8590 0.5692797 Africa
#> Angola Angola -376.5048 0.2093399 Africa
#> Argentina Argentina -389.6063 0.2317084 Americas
#> Australia Australia -376.1163 0.2277238 Oceania</code></pre>
</div>
<div class="col-md-6">
<p><strong>tidyverse</strong></p>
<pre class="r"><code>nested_df %>%
mutate(coefs = map(fit, coef),
intercept = map_dbl(coefs, 1),
slope = map_dbl(coefs, 2)) %>%
select(country, continent, intercept, slope)
#> # A tibble: 142 x 4
#> # Groups: country, continent [710]
#> country continent intercept slope
#> <fct> <fct> <dbl> <dbl>
#> 1 Afghanistan Asia -508. 0.275
#> 2 Albania Europe -594. 0.335
#> 3 Algeria Africa -1068. 0.569
#> 4 Angola Africa -377. 0.209
#> 5 Argentina Americas -390. 0.232
#> 6 Australia Oceania -376. 0.228
#> 7 Austria Europe -406. 0.242
#> 8 Bahrain Asia -860. 0.468
#> 9 Bangladesh Asia -936. 0.498
#> 10 Belgium Europe -340. 0.209
#> # … with 132 more rows</code></pre>
</div>
</div>
<p>This illustrates the post-processing that is often necessary in a base workflow. We need to restore the country info from the names of <code>by_obj</code>, use them to look up the continents in gapminder, and then restore the original factor levels for both country and continent. It illustrates the payoff for temporarily tolerating the <code>data</code> and <code>fit</code> list-columns in the nested data frame used in the tidyverse workflow. The country and continent factors remain intact and directly associated with the data and fits.</p>
</div>
<i class="fab fa-creative-commons fa-2x"></i><i class="fab fa-creative-commons-by fa-2x"></i><i class="fab fa-creative-commons-sa fa-2x"></i>
</div>
</div>
</div>
<script>
// add bootstrap table styles to pandoc tables
function bootstrapStylePandocTables() {
$('tr.header').parent('thead').parent('table').addClass('table table-condensed');
}
$(document).ready(function () {
bootstrapStylePandocTables();
});
</script>
<!-- tabsets -->
<script>
$(document).ready(function () {
window.buildTabsets("TOC");
});
$(document).ready(function () {
$('.tabset-dropdown > .nav-tabs > li').click(function () {
$(this).parent().toggleClass('nav-tabs-open')
});
});
</script>
<!-- code folding -->
<script>
$(document).ready(function () {
// move toc-ignore selectors from section div to header
$('div.section.toc-ignore')
.removeClass('toc-ignore')
.children('h1,h2,h3,h4,h5').addClass('toc-ignore');
// establish options
var options = {
selectors: "h1,h2,h3",
theme: "bootstrap3",
context: '.toc-content',
hashGenerator: function (text) {
return text.replace(/[.\\/?&!#<>]/g, '').replace(/\s/g, '_').toLowerCase();
},
ignoreSelector: ".toc-ignore",
scrollTo: 0
};
options.showAndHide = true;
options.smoothScroll = true;
// tocify
var toc = $("#TOC").tocify(options).data("toc-tocify");
});
</script>
<!-- dynamically load mathjax for compatibility with self-contained -->
<script>
(function () {
var script = document.createElement("script");
script.type = "text/javascript";
script.src = "https://mathjax.rstudio.com/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML";
document.getElementsByTagName("head")[0].appendChild(script);
})();
</script>
</body>
</html>