-
Notifications
You must be signed in to change notification settings - Fork 1
/
three-reasons-a-data-engineer-should-learn-scala.html
310 lines (268 loc) · 21.2 KB
/
three-reasons-a-data-engineer-should-learn-scala.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
<!DOCTYPE html>
<html lang="en" prefix="og: http://ogp.me/ns# fb: https://www.facebook.com/2008/fbml">
<head>
<title>Three Reasons a Data Engineer Should Learn Scala - Stackdiver as a Service</title>
<!-- Using the latest rendering mode for IE -->
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<link rel="canonical" href="https://www.lyh.me/three-reasons-a-data-engineer-should-learn-scala.html">
<meta name="author" content="Neville Li" />
<meta name="keywords" content="data,scala" />
<meta name="description" content="This article was written in collaboration with Hakka Labs (original link) There has been a lot of debate over Scala lately, including criticisms like this, this, this, and defenses like this and this. Most of the criticisms seem to focus on the language’s complexity, performance, and integration with existing tools and libraries, while some praise its elegant syntax, powerful type system, and good fit for domain-specific languages. However most of the discussions seem based on experiences building production backend or web systems where there are a lot of other options already. There are mature, battle tested options like Java, Erlang or even PHP, and there are Go, node.js, or Python for those who are more adventurous or prefer agility over performance. Here I want to argue that there’s a best tool for every job, and Scala shines for data processing and machine learning, for the following reasons: good balance between productivity and performance integration with big data ecosystem functional paradigm Productivity without sacrificing performance In the big data & machine learning world where most developers are from Python/R/Matlab background, Scala’s syntax, or the subset needed for the domain, is a lot less intimidating than that …" />
<meta property="og:site_name" content="Stackdiver as a Service" />
<meta property="og:type" content="article"/>
<meta property="og:title" content="Three Reasons a Data Engineer Should Learn Scala"/>
<meta property="og:url" content="https://www.lyh.me/three-reasons-a-data-engineer-should-learn-scala.html"/>
<meta property="og:description" content="This article was written in collaboration with Hakka Labs (original link) There has been a lot of debate over Scala lately, including criticisms like this, this, this, and defenses like this and this. Most of the criticisms seem to focus on the language’s complexity, performance, and integration with existing tools and libraries, while some praise its elegant syntax, powerful type system, and good fit for domain-specific languages. However most of the discussions seem based on experiences building production backend or web systems where there are a lot of other options already. There are mature, battle tested options like Java, Erlang or even PHP, and there are Go, node.js, or Python for those who are more adventurous or prefer agility over performance. Here I want to argue that there’s a best tool for every job, and Scala shines for data processing and machine learning, for the following reasons: good balance between productivity and performance integration with big data ecosystem functional paradigm Productivity without sacrificing performance In the big data & machine learning world where most developers are from Python/R/Matlab background, Scala’s syntax, or the subset needed for the domain, is a lot less intimidating than that …"/>
<meta property="article:published_time" content="2014-11-17" />
<meta property="article:section" content="code" />
<meta property="article:tag" content="data" />
<meta property="article:tag" content="scala" />
<meta property="article:author" content="Neville Li" />
<!-- Bootstrap -->
<link rel="stylesheet" href="https://www.lyh.me/theme/css/bootstrap.min.css" type="text/css"/>
<link href="https://www.lyh.me/theme/css/font-awesome.min.css" rel="stylesheet">
<link href="https://www.lyh.me/theme/css/pygments/monokai.css" rel="stylesheet">
<link href="https://www.lyh.me/theme/css/typogrify.css" rel="stylesheet">
<link rel="stylesheet" href="https://www.lyh.me/theme/css/style.css" type="text/css"/>
<link href="https://www.lyh.me/feeds/all.atom.xml" type="application/atom+xml" rel="alternate"
title="Stackdiver as a Service ATOM Feed"/>
<link href="https://www.lyh.me/feeds/code.atom.xml" type="application/atom+xml" rel="alternate"
title="Stackdiver as a Service code ATOM Feed"/>
</head>
<body>
<div class="navbar navbar-default navbar-fixed-top" role="navigation">
<div class="container">
<div class="navbar-header">
<button type="button" class="navbar-toggle" data-toggle="collapse" data-target=".navbar-ex1-collapse">
<span class="sr-only">Toggle navigation</span>
<span class="icon-bar"></span>
<span class="icon-bar"></span>
<span class="icon-bar"></span>
</button>
<a href="https://www.lyh.me/" class="navbar-brand">
Stackdiver as a Service </a>
</div>
<div class="collapse navbar-collapse navbar-ex1-collapse">
<ul class="nav navbar-nav">
<li><a href="https://www.lyh.me/pages/about-me.html">
About Me
</a></li>
</ul>
<ul class="nav navbar-nav navbar-right">
</ul>
</div>
<!-- /.navbar-collapse -->
</div>
</div> <!-- /.navbar -->
<!-- Banner -->
<!-- End Banner -->
<!-- Content Container -->
<div class="container">
<div class="row">
<div class="col-sm-9">
<section id="content">
<article>
<header class="page-header">
<h1>
<a href="https://www.lyh.me/three-reasons-a-data-engineer-should-learn-scala.html"
rel="bookmark"
title="Permalink to Three Reasons a Data Engineer Should Learn Scala">
Three Reasons a Data Engineer Should Learn Scala
</a>
</h1>
</header>
<div class="entry-content">
<div class="panel">
<div class="panel-body">
<footer class="post-info">
<span class="label label-default">Date</span>
<span class="published">
<i class="fa fa-calendar"></i><time datetime="2014-11-17T23:15:00-05:00"> Mon 17 November 2014</time>
</span>
<span class="label label-default">Category</span>
<a href="https://www.lyh.me/category/code.html">code</a>
<span class="label label-default">Tags</span>
<a href="https://www.lyh.me/tag/data.html">data</a>
/
<a href="https://www.lyh.me/tag/scala.html">scala</a>
</footer><!-- /.post-info --> </div>
</div>
<p><em>This article was written in collaboration with <a href="https://www.hakkalabs.co">Hakka Labs</a> (<a href="https://www.hakkalabs.co/articles/three-reasons-data-eng-learn-scala">original link</a>)</em></p>
<p>There has been a lot of debate over Scala lately, including criticisms like <a href="http://java.dzone.com/articles/i-dont-scala">this</a>, <a href="http://overwatering.org/blog/2013/12/scala-1-star-would-not-program-again/">this</a>, <a href="http://www.infoq.com/news/2011/11/yammer-scala">this</a>, and defenses like <a href="http://blog.gridgainsystems.com/in-defense-of-scala-response-to-i-dont-like-scala/">this</a> and <a href="http://blog.gridgainsystems.com/in-defense-of-scala-part-2/">this</a>. Most of the criticisms seem to focus on the language’s complexity, performance, and integration with existing tools and libraries, while some praise its elegant syntax, powerful type system, and good fit for domain-specific languages.</p>
<p>However most of the discussions seem based on experiences building production backend or web systems where there are a lot of other options already. There are mature, battle tested options like Java, Erlang or even <span class="caps">PHP</span>, and there are Go, node.js, or Python for those who are more adventurous or prefer agility over performance.</p>
<p>Here I want to argue that there’s a best tool for every job, and Scala shines for data processing and machine learning, for the following reasons:</p>
<ul>
<li>good balance between productivity and performance</li>
<li>integration with big data ecosystem</li>
<li>functional paradigm</li>
</ul>
<h2>Productivity without sacrificing performance</h2>
<p>In the big data <span class="amp">&</span> machine learning world where most developers are from Python/R/Matlab background, Scala’s syntax, or the subset needed for the domain, is a lot less intimidating than that of Java or C++. In my experience, basic syntax collections <span class="caps">API</span> and lambda (about 20% of the language features) is all that’s needed for a new hire with no prior experience to become productive in processing data. Libraries like <a href="https://github.com/scalanlp/breeze">Breeze</a>, <a href="https://code.google.com/p/scalalab/">ScalaLab</a> and <a href="https://github.com/BIDData/BIDMach">BIDMach</a> mimic syntax of popular tools with operator overloading and other syntactic sugar which are otherwise impossible in many mainstream languages. At the same time, performance is usually better than traditional tools like Python or R. As one’s skill develops over time, there’s a clear transition path from imperative to more elegant <span class="caps">FP</span> style code while maintaining or even improving performance.</p>
<p>The chart below clearly shows the growth of our main <a href="https://github.com/twitter/scalding">scalding</a> repository over time. We spent a few months experimenting and after doubling the team size in early 2014, the number of jobs and <span class="caps">LOC</span> also exploded. Most developers have little Scala or even Java experience and some of the jobs are doing complex machine learning stuff. </p>
<p><img alt="rec-sys-scalding" src="https://www.lyh.me/images/rec-sys-scalding.jpg"></p>
<h2>Ecosystem</h2>
<p>Scala also integrates well with the big data eco-system, which is mostly <span class="caps">JVM</span> based. There are frameworks on top of Java libraries like Scalding (Cascading), <a href="https://github.com/twitter/summingbird">Summingbird</a> (Scalding and Storm), <a href="http://crunch.apache.org/scrunch.html">Scrunch</a> (Crunch), <a href="http://flink.incubator.apache.org/">Flink</a> (Java core with Scala <span class="caps">API</span>), and ones built from scratch but interface with <span class="caps">JVM</span> systems, like <a href="http://spark.apache.org/">Spark</a> and <a href="http://kafka.apache.org/">Kafka</a>. The Scala APIs are usually more flexible than say Hadoop streaming with Python/Perl, PySpark or Python/Ruby bolts in Storm, since you have direct access to the underlying <span class="caps">API</span>. There are also a wide range of data storage solutions that are built for or work well with <span class="caps">JVM</span> like <a href="http://cassandra.apache.org/">Cassandra</a>, <a href="http://hbase.apache.org/">HBase</a>, <a href="http://www.project-voldemort.com/voldemort/">Voldemort</a> and <a href="http://www.datomic.com/">Datomic</a>.</p>
<h2>Functional paradigm</h2>
<p>A third benefit is the functional paradigm which fits well within the Map/Reduce and big data model. Batch processing works on top of immutable data, transforms with map and reduce operations, and generates new copies. Real time log streams are essentially lazy streams. Most Scala data frameworks have the notion of some abstract data type that’s extremely consistent with Scala’s collection <span class="caps">API</span>. A glance at <a href="http://twitter.github.io/scalding/com/twitter/scalding/typed/TypedPipe.html">TypedPipe</a> in Scalding and <a href="http://spark.apache.org/docs/latest/api/scala/index.html#org.apache.spark.rdd.RDD"><span class="caps">RDD</span></a> in Spark, and you’ll see that they all have the same set of methods, e.g. map, flatMap, filter, reduce, fold and groupBy. One could just learn the standard collection and easily pick up one of the libraries. Many libraries also have frequent reference of category theory, more specifically properties of semigroup, monoid, and group to guarantee the correctness of distributed operations. Equipped such knowledge it’ll be a lot easier to understand techniques like map-side reduce.</p>
<p>With these benefits I would say Scala is here to stay in the big data world and there are few contenders close in the competition.</p>
</div>
<!-- /.entry-content -->
<hr />
<!-- AddThis Button BEGIN -->
<div class="addthis_toolbox addthis_default_style">
<a class="addthis_button_facebook_like" fb:like:layout="button_count"></a>
<a class="addthis_button_tweet"></a>
<a class="addthis_button_google_plusone" g:plusone:size="medium"></a>
</div>
<!-- AddThis Button END -->
<hr/>
<section class="comments" id="comments">
<h2>Comments</h2>
<div id="disqus_thread"></div>
<script type="text/javascript">
/* * * CONFIGURATION VARIABLES: EDIT BEFORE PASTING INTO YOUR WEBPAGE * * */
var disqus_shortname = 'lyh'; // required: replace example with your forum shortname
var disqus_config = function () {
this.language = "en";
this.page.identifier = '2014-11-17-three-reasons-a-data-engineer-should-learn-scala';
this.page.url = 'https://www.lyh.me/three-reasons-a-data-engineer-should-learn-scala.html';
};
/* * * DON'T EDIT BELOW THIS LINE * * */
(function () {
var dsq = document.createElement('script');
dsq.type = 'text/javascript';
dsq.async = true;
dsq.src = '//' + disqus_shortname + '.disqus.com/embed.js';
(document.getElementsByTagName('head')[0] || document.getElementsByTagName('body')[0]).appendChild(dsq);
})();
</script>
<noscript>Please enable JavaScript to view the <a href="http://disqus.com/?ref_noscript">comments powered by
Disqus.</a></noscript>
<a href="http://disqus.com" class="dsq-brlink">comments powered by <span class="logo-disqus">Disqus</span></a>
</section>
</article>
</section>
</div>
<div class="col-sm-3" id="sidebar">
<aside>
<div id="aboutme">
<p>
<img width="100%" class="img-thumbnail" src="https://www.lyh.me//avatar.jpg"/>
</p>
<p>
<strong>About Neville Li</strong><br/>
Data infrastructure @<a href="https://twitter.com/Spotify">Spotify</a>, ex-@<a href="https://twitter.com/Yahoo">Yahoo</a> search, creator of <a href="https://github.com/spotify/scio">Scio</a>, technical cave & wreck diver, lefty guitar player
</p>
</div><!-- Sidebar -->
<section class="well well-sm">
<ul class="list-group list-group-flush">
<!-- Sidebar/Social -->
<li class="list-group-item">
<h4><i class="fa fa-home fa-lg"></i><span class="icon-label">Social</span></h4>
<ul class="list-group" id="social">
<li class="list-group-item"><a href="https://open.spotify.com/user/sinisa_lyh"><i class="fa fa-spotify fa-lg"></i> Spotify</a></li>
<li class="list-group-item"><a href="https://github.com/nevillelyh"><i class="fa fa-github-square fa-lg"></i> GitHub</a></li>
<li class="list-group-item"><a href="https://twitter.com/sinisa_lyh"><i class="fa fa-twitter-square fa-lg"></i> Twitter</a></li>
<li class="list-group-item"><a href="https://www.slideshare.net/sinisalyh"><i class="fa fa-slideshare fa-lg"></i> SlideShare</a></li>
<li class="list-group-item"><a href="https://www.youtube.com/user/sinisalyh/videos"><i class="fa fa-youtube-square fa-lg"></i> YouTube</a></li>
<li class="list-group-item"><a href="https://www.instagram.com/sinisa/"><i class="fa fa-instagram fa-lg"></i> Instagram</a></li>
<li class="list-group-item"><a href="https://www.flickr.com/photos/sinisa_lyh"><i class="fa fa-flickr fa-lg"></i> Flickr</a></li>
</ul>
</li>
<!-- End Sidebar/Social -->
<!-- Sidebar/Recent Posts -->
<li class="list-group-item">
<h4><i class="fa fa-home fa-lg"></i><span class="icon-label">Recent Posts</span></h4>
<ul class="list-group" id="recentposts">
<li class="list-group-item"><a href="https://www.lyh.me/magnolify.html">Magnolify</a></li>
<li class="list-group-item"><a href="https://www.lyh.me/featran.html">Featran</a></li>
<li class="list-group-item"><a href="https://www.lyh.me/automatic-type-class-derivation-with-shapeless.html">Automatic type-class derivation with Shapeless</a></li>
<li class="list-group-item"><a href="https://www.lyh.me/lambda-serialization.html">Lambda serialization</a></li>
<li class="list-group-item"><a href="https://www.lyh.me/lawfulness-of-aggregatebykey.html">Lawfulness of aggregateByKey</a></li>
</ul>
</li>
<!-- End Sidebar/Recent Posts -->
<!-- Sidebar/Categories -->
<li class="list-group-item">
<h4><i class="fa fa-home fa-lg"></i><span class="icon-label">Categories</span></h4>
<ul class="list-group" id="categories">
<li class="list-group-item">
<a href="https://www.lyh.me/category/code.html"><i class="fa fa-folder-open fa-lg"></i>code</a>
</li>
<li class="list-group-item">
<a href="https://www.lyh.me/category/misc.html"><i class="fa fa-folder-open fa-lg"></i>misc</a>
</li>
</ul>
</li>
<!-- End Sidebar/Categories -->
<!-- Sidebar/Twitter Timeline -->
<li class="list-group-item">
<h4><i class="fa fa-twitter fa-lg"></i><span class="icon-label">Latest Tweets</span></h4>
<div id="twitter_timeline">
<a class="twitter-timeline" data-width="250" data-height="300" data-dnt="true" data-theme="light" href="https://twitter.com/sinisa_lyh">Tweets by sinisa_lyh</a> <script async src="//platform.twitter.com/widgets.js" charset="utf-8"></script>
</div>
</li>
<!-- End Sidebar/Twitter Timeline -->
</ul>
</section>
<!-- End Sidebar --> </aside>
</div>
</div>
</div>
<!-- End Content Container -->
<footer>
<div class="container">
<hr>
<div class="row">
<div class="col-xs-10">© 2020 Neville Li
· Powered by <a href="https://github.com/getpelican/pelican-themes/tree/master/pelican-bootstrap3" target="_blank">pelican-bootstrap3</a>,
<a href="http://docs.getpelican.com/" target="_blank">Pelican</a>,
<a href="http://getbootstrap.com" target="_blank">Bootstrap</a> <p><small> <a rel="license" href="https://creativecommons.org/licenses/by-nc/4.0/deed.en"><img alt="Creative Commons License" style="border-width:0" src="//i.creativecommons.org/l/by-nc/4.0/80x15.png" /></a>
Content
licensed under a <a rel="license" href="https://creativecommons.org/licenses/by-nc/4.0/deed.en">Creative Commons Attribution-NonCommercial 4.0 International License</a>, except where indicated otherwise.
</small></p>
</div>
<div class="col-xs-2"><p class="pull-right"><i class="fa fa-arrow-up"></i> <a href="#">Back to top</a></p></div>
</div>
</div>
</footer>
<script src="https://www.lyh.me/theme/js/jquery.min.js"></script>
<!-- Include all compiled plugins (below), or include individual files as needed -->
<script src="https://www.lyh.me/theme/js/bootstrap.min.js"></script>
<!-- Enable responsive features in IE8 with Respond.js (https://github.com/scottjehl/Respond) -->
<script src="https://www.lyh.me/theme/js/respond.min.js"></script>
<!-- Disqus -->
<script type="text/javascript">
/* * * CONFIGURATION VARIABLES: EDIT BEFORE PASTING INTO YOUR WEBPAGE * * */
var disqus_shortname = 'lyh'; // required: replace example with your forum shortname
/* * * DON'T EDIT BELOW THIS LINE * * */
(function () {
var s = document.createElement('script');
s.async = true;
s.type = 'text/javascript';
s.src = '//' + disqus_shortname + '.disqus.com/count.js';
(document.getElementsByTagName('HEAD')[0] || document.getElementsByTagName('BODY')[0]).appendChild(s);
}());
</script>
<!-- End Disqus Code -->
<!-- Google Analytics -->
<script type="text/javascript">
var _gaq = _gaq || [];
_gaq.push(['_setAccount', 'UA-6988688-5']);
_gaq.push(['_trackPageview']);
(function () {
var ga = document.createElement('script');
ga.type = 'text/javascript';
ga.async = true;
ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
var s = document.getElementsByTagName('script')[0];
s.parentNode.insertBefore(ga, s);
})();
</script>
<!-- End Google Analytics Code -->
<script type="text/javascript">var addthis_config = {"data_track_addressbar": true};</script>
<script type="text/javascript" src="//s7.addthis.com/js/300/addthis_widget.js#pubid=sinisalyh"></script>
</body>
</html>