Google Git
Sign in
apache / spark-website / 2c562896505588421cce257d111c8ba3805528c2 / . / site / docs / 1.6.2 / sparkr.html
blob: cd0c190e93fd8b77707314d3703f7fa21d312842 [file] [log] [blame]
<!DOCTYPE html>
<!--[if lt IE 7]> <html class="no-js lt-ie9 lt-ie8 lt-ie7"> <![endif]-->
<!--[if IE 7]> <html class="no-js lt-ie9 lt-ie8"> <![endif]-->
<!--[if IE 8]> <html class="no-js lt-ie9"> <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js"> <!--<![endif]-->
<head>
<meta charset="utf-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge,chrome=1">
<title>SparkR (R on Spark) - Spark 1.6.2 Documentation</title>
<link rel="stylesheet" href="css/bootstrap.min.css">
<style>
body {
padding-top: 60px;
padding-bottom: 40px;
}
</style>
<meta name="viewport" content="width=device-width">
<link rel="stylesheet" href="css/bootstrap-responsive.min.css">
<link rel="stylesheet" href="css/main.css">
<script src="js/vendor/modernizr-2.6.1-respond-1.1.0.min.js"></script>
<link rel="stylesheet" href="css/pygments-default.css">
<!-- Google analytics script -->
<script type="text/javascript">
var _gaq = _gaq || [];
_gaq.push(['_setAccount', 'UA-32518208-2']);
_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>
</head>
<body>
<!--[if lt IE 7]>
<p class="chromeframe">You are using an outdated browser. <a href="http://browsehappy.com/">Upgrade your browser today</a> or <a href="http://www.google.com/chromeframe/?redirect=true">install Google Chrome Frame</a> to better experience this site.</p>
<![endif]-->
<!-- This code is taken from http://twitter.github.com/bootstrap/examples/hero.html -->
<div class="navbar navbar-fixed-top" id="topbar">
<div class="navbar-inner">
<div class="container">
<div class="brand"><a href="index.html">
<img src="img/spark-logo-hd.png" style="height:50px;"/></a><span class="version">1.6.2</span>
</div>
<ul class="nav">
<!--TODO(andyk): Add class="active" attribute to li some how.-->
<li><a href="index.html">Overview</a></li>
<li class="dropdown">
<a href="#" class="dropdown-toggle" data-toggle="dropdown">Programming Guides<b class="caret"></b></a>
<ul class="dropdown-menu">
<li><a href="quick-start.html">Quick Start</a></li>
<li><a href="programming-guide.html">Spark Programming Guide</a></li>
<li class="divider"></li>
<li><a href="streaming-programming-guide.html">Spark Streaming</a></li>
<li><a href="sql-programming-guide.html">DataFrames, Datasets and SQL</a></li>
<li><a href="mllib-guide.html">MLlib (Machine Learning)</a></li>
<li><a href="graphx-programming-guide.html">GraphX (Graph Processing)</a></li>
<li><a href="bagel-programming-guide.html">Bagel (Pregel on Spark)</a></li>
<li><a href="sparkr.html">SparkR (R on Spark)</a></li>
</ul>
</li>
<li class="dropdown">
<a href="#" class="dropdown-toggle" data-toggle="dropdown">API Docs<b class="caret"></b></a>
<ul class="dropdown-menu">
<li><a href="api/scala/index.html#org.apache.spark.package">Scala</a></li>
<li><a href="api/java/index.html">Java</a></li>
<li><a href="api/python/index.html">Python</a></li>
<li><a href="api/R/index.html">R</a></li>
</ul>
</li>
<li class="dropdown">
<a href="#" class="dropdown-toggle" data-toggle="dropdown">Deploying<b class="caret"></b></a>
<ul class="dropdown-menu">
<li><a href="cluster-overview.html">Overview</a></li>
<li><a href="submitting-applications.html">Submitting Applications</a></li>
<li class="divider"></li>
<li><a href="spark-standalone.html">Spark Standalone</a></li>
<li><a href="running-on-mesos.html">Mesos</a></li>
<li><a href="running-on-yarn.html">YARN</a></li>
<li class="divider"></li>
<li><a href="ec2-scripts.html">Amazon EC2</a></li>
</ul>
</li>
<li class="dropdown">
<a href="api.html" class="dropdown-toggle" data-toggle="dropdown">More<b class="caret"></b></a>
<ul class="dropdown-menu">
<li><a href="configuration.html">Configuration</a></li>
<li><a href="monitoring.html">Monitoring</a></li>
<li><a href="tuning.html">Tuning Guide</a></li>
<li><a href="job-scheduling.html">Job Scheduling</a></li>
<li><a href="security.html">Security</a></li>
<li><a href="hardware-provisioning.html">Hardware Provisioning</a></li>
<li class="divider"></li>
<li><a href="building-spark.html">Building Spark</a></li>
<li><a href="https://cwiki.apache.org/confluence/display/SPARK/Contributing+to+Spark">Contributing to Spark</a></li>
<li><a href="https://cwiki.apache.org/confluence/display/SPARK/Supplemental+Spark+Projects">Supplemental Projects</a></li>
</ul>
</li>
</ul>
<!--<p class="navbar-text pull-right"><span class="version-text">v1.6.2</span></p>-->
</div>
</div>
</div>
<div class="container-wrapper">
<div class="content" id="content">
<h1 class="title">SparkR (R on Spark)</h1>
<ul id="markdown-toc">
<li><a href="#overview" id="markdown-toc-overview">Overview</a></li>
<li><a href="#sparkr-dataframes" id="markdown-toc-sparkr-dataframes">SparkR DataFrames</a> <ul>
<li><a href="#starting-up-sparkcontext-sqlcontext" id="markdown-toc-starting-up-sparkcontext-sqlcontext">Starting Up: SparkContext, SQLContext</a></li>
<li><a href="#starting-up-from-rstudio" id="markdown-toc-starting-up-from-rstudio">Starting Up from RStudio</a></li>
<li><a href="#creating-dataframes" id="markdown-toc-creating-dataframes">Creating DataFrames</a> <ul>
<li><a href="#from-local-data-frames" id="markdown-toc-from-local-data-frames">From local data frames</a></li>
<li><a href="#from-data-sources" id="markdown-toc-from-data-sources">From Data Sources</a></li>
<li><a href="#from-hive-tables" id="markdown-toc-from-hive-tables">From Hive tables</a></li>
</ul>
</li>
<li><a href="#dataframe-operations" id="markdown-toc-dataframe-operations">DataFrame Operations</a> <ul>
<li><a href="#selecting-rows-columns" id="markdown-toc-selecting-rows-columns">Selecting rows, columns</a></li>
<li><a href="#grouping-aggregation" id="markdown-toc-grouping-aggregation">Grouping, Aggregation</a></li>
<li><a href="#operating-on-columns" id="markdown-toc-operating-on-columns">Operating on Columns</a></li>
</ul>
</li>
<li><a href="#running-sql-queries-from-sparkr" id="markdown-toc-running-sql-queries-from-sparkr">Running SQL Queries from SparkR</a></li>
</ul>
</li>
<li><a href="#machine-learning" id="markdown-toc-machine-learning">Machine Learning</a> <ul>
<li><a href="#gaussian-glm-model" id="markdown-toc-gaussian-glm-model">Gaussian GLM model</a></li>
<li><a href="#binomial-glm-model" id="markdown-toc-binomial-glm-model">Binomial GLM model</a></li>
</ul>
</li>
<li><a href="#r-function-name-conflicts" id="markdown-toc-r-function-name-conflicts">R Function Name Conflicts</a></li>
<li><a href="#migration-guide" id="markdown-toc-migration-guide">Migration Guide</a> <ul>
<li><a href="#upgrading-from-sparkr-15x-to-16" id="markdown-toc-upgrading-from-sparkr-15x-to-16">Upgrading From SparkR 1.5.x to 1.6</a></li>
</ul>
</li>
</ul>
<h1 id="overview">Overview</h1>
<p>SparkR is an R package that provides a light-weight frontend to use Apache Spark from R.
In Spark 1.6.2, SparkR provides a distributed data frame implementation that
supports operations like selection, filtering, aggregation etc. (similar to R data frames,
<a href="https://github.com/hadley/dplyr">dplyr</a>) but on large datasets. SparkR also supports distributed
machine learning using MLlib.</p>
<h1 id="sparkr-dataframes">SparkR DataFrames</h1>
<p>A DataFrame is a distributed collection of data organized into named columns. It is conceptually
equivalent to a table in a relational database or a data frame in R, but with richer
optimizations under the hood. DataFrames can be constructed from a wide array of sources such as:
structured data files, tables in Hive, external databases, or existing local R data frames.</p>
<p>All of the examples on this page use sample data included in R or the Spark distribution and can be run using the <code>./bin/sparkR</code> shell.</p>
<h2 id="starting-up-sparkcontext-sqlcontext">Starting Up: SparkContext, SQLContext</h2>
<div data-lang="r">
<p>The entry point into SparkR is the <code>SparkContext</code> which connects your R program to a Spark cluster.
You can create a <code>SparkContext</code> using <code>sparkR.init</code> and pass in options such as the application name
, any spark packages depended on, etc. Further, to work with DataFrames we will need a <code>SQLContext</code>,
which can be created from the SparkContext. If you are working from the <code>sparkR</code> shell, the
<code>SQLContext</code> and <code>SparkContext</code> should already be created for you, and you would not need to call
<code>sparkR.init</code>.</p>
<div data-lang="r">
<div class="highlight"><pre><code class="language-r" data-lang="r">sc <span class="o">&lt;-</span> sparkR.init<span class="p">()</span>
sqlContext <span class="o">&lt;-</span> sparkRSQL.init<span class="p">(</span>sc<span class="p">)</span></code></pre></div>
</div>
<h2 id="starting-up-from-rstudio">Starting Up from RStudio</h2>
<p>You can also start SparkR from RStudio. You can connect your R program to a Spark cluster from
RStudio, R shell, Rscript or other R IDEs. To start, make sure SPARK_HOME is set in environment
(you can check <a href="https://stat.ethz.ch/R-manual/R-devel/library/base/html/Sys.getenv.html">Sys.getenv</a>),
load the SparkR package, and call <code>sparkR.init</code> as below. In addition to calling <code>sparkR.init</code>, you
could also specify certain Spark driver properties. Normally these
<a href="configuration.html#application-properties">Application properties</a> and
<a href="configuration.html#runtime-environment">Runtime Environment</a> cannot be set programmatically, as the
driver JVM process would have been started, in this case SparkR takes care of this for you. To set
them, pass them as you would other configuration properties in the <code>sparkEnvir</code> argument to
<code>sparkR.init()</code>.</p>
<div data-lang="r">
<div class="highlight"><pre><code class="language-r" data-lang="r"><span class="kr">if</span> <span class="p">(</span><span class="kp">nchar</span><span class="p">(</span><span class="kp">Sys.getenv</span><span class="p">(</span><span class="s">&quot;SPARK_HOME&quot;</span><span class="p">))</span> <span class="o">&lt;</span> <span class="m">1</span><span class="p">)</span> <span class="p">{</span>
<span class="kp">Sys.setenv</span><span class="p">(</span>SPARK_HOME <span class="o">=</span> <span class="s">&quot;/home/spark&quot;</span><span class="p">)</span>
<span class="p">}</span>
<span class="kn">library</span><span class="p">(</span>SparkR<span class="p">,</span> lib.loc <span class="o">=</span> <span class="kt">c</span><span class="p">(</span><span class="kp">file.path</span><span class="p">(</span><span class="kp">Sys.getenv</span><span class="p">(</span><span class="s">&quot;SPARK_HOME&quot;</span><span class="p">),</span> <span class="s">&quot;R&quot;</span><span class="p">,</span> <span class="s">&quot;lib&quot;</span><span class="p">)))</span>
sc <span class="o">&lt;-</span> sparkR.init<span class="p">(</span>master <span class="o">=</span> <span class="s">&quot;local[*]&quot;</span><span class="p">,</span> sparkEnvir <span class="o">=</span> <span class="kt">list</span><span class="p">(</span>spark.driver.memory<span class="o">=</span><span class="s">&quot;2g&quot;</span><span class="p">))</span></code></pre></div>
</div>
<p>The following options can be set in <code>sparkEnvir</code> with <code>sparkR.init</code> from RStudio:</p>
<table class="table">
<tr><th>Property Name</th><th>Property group</th><th><code>spark-submit</code> equivalent</th></tr>
<tr>
<td><code>spark.driver.memory</code></td>
<td>Application Properties</td>
<td><code>--driver-memory</code></td>
</tr>
<tr>
<td><code>spark.driver.extraClassPath</code></td>
<td>Runtime Environment</td>
<td><code>--driver-class-path</code></td>
</tr>
<tr>
<td><code>spark.driver.extraJavaOptions</code></td>
<td>Runtime Environment</td>
<td><code>--driver-java-options</code></td>
</tr>
<tr>
<td><code>spark.driver.extraLibraryPath</code></td>
<td>Runtime Environment</td>
<td><code>--driver-library-path</code></td>
</tr>
</table>
</div>
<h2 id="creating-dataframes">Creating DataFrames</h2>
<p>With a <code>SQLContext</code>, applications can create <code>DataFrame</code>s from a local R data frame, from a <a href="sql-programming-guide.html#hive-tables">Hive table</a>, or from other <a href="sql-programming-guide.html#data-sources">data sources</a>.</p>
<h3 id="from-local-data-frames">From local data frames</h3>
<p>The simplest way to create a data frame is to convert a local R data frame into a SparkR DataFrame. Specifically we can use <code>createDataFrame</code> and pass in the local R data frame to create a SparkR DataFrame. As an example, the following creates a <code>DataFrame</code> based using the <code>faithful</code> dataset from R.</p>
<div data-lang="r">
<div class="highlight"><pre><code class="language-r" data-lang="r">df <span class="o">&lt;-</span> createDataFrame<span class="p">(</span>sqlContext<span class="p">,</span> faithful<span class="p">)</span>
<span class="c1"># Displays the content of the DataFrame to stdout</span>
<span class="kp">head</span><span class="p">(</span>df<span class="p">)</span>
<span class="c1">## eruptions waiting</span>
<span class="c1">##1 3.600 79</span>
<span class="c1">##2 1.800 54</span>
<span class="c1">##3 3.333 74</span></code></pre></div>
</div>
<h3 id="from-data-sources">From Data Sources</h3>
<p>SparkR supports operating on a variety of data sources through the <code>DataFrame</code> interface. This section describes the general methods for loading and saving data using Data Sources. You can check the Spark SQL programming guide for more <a href="sql-programming-guide.html#manually-specifying-options">specific options</a> that are available for the built-in data sources.</p>
<p>The general method for creating DataFrames from data sources is <code>read.df</code>. This method takes in the <code>SQLContext</code>, the path for the file to load and the type of data source. SparkR supports reading JSON and Parquet files natively and through <a href="http://spark-packages.org/">Spark Packages</a> you can find data source connectors for popular file formats like <a href="http://spark-packages.org/package/databricks/spark-csv">CSV</a> and <a href="http://spark-packages.org/package/databricks/spark-avro">Avro</a>. These packages can either be added by
specifying <code>--packages</code> with <code>spark-submit</code> or <code>sparkR</code> commands, or if creating context through <code>init</code>
you can specify the packages with the <code>packages</code> argument.</p>
<div data-lang="r">
<div class="highlight"><pre><code class="language-r" data-lang="r">sc <span class="o">&lt;-</span> sparkR.init<span class="p">(</span>sparkPackages<span class="o">=</span><span class="s">&quot;com.databricks:spark-csv_2.11:1.0.3&quot;</span><span class="p">)</span>
sqlContext <span class="o">&lt;-</span> sparkRSQL.init<span class="p">(</span>sc<span class="p">)</span></code></pre></div>
</div>
<p>We can see how to use data sources using an example JSON input file. Note that the file that is used here is <em>not</em> a typical JSON file. Each line in the file must contain a separate, self-contained valid JSON object. As a consequence, a regular multi-line JSON file will most often fail.</p>
<div data-lang="r">
<div class="highlight"><pre><code class="language-r" data-lang="r">people <span class="o">&lt;-</span> read.df<span class="p">(</span>sqlContext<span class="p">,</span> <span class="s">&quot;./examples/src/main/resources/people.json&quot;</span><span class="p">,</span> <span class="s">&quot;json&quot;</span><span class="p">)</span>
<span class="kp">head</span><span class="p">(</span>people<span class="p">)</span>
<span class="c1">## age name</span>
<span class="c1">##1 NA Michael</span>
<span class="c1">##2 30 Andy</span>
<span class="c1">##3 19 Justin</span>
<span class="c1"># SparkR automatically infers the schema from the JSON file</span>
printSchema<span class="p">(</span>people<span class="p">)</span>
<span class="c1"># root</span>
<span class="c1"># |-- age: integer (nullable = true)</span>
<span class="c1"># |-- name: string (nullable = true)</span></code></pre></div>
</div>
<p>The data sources API can also be used to save out DataFrames into multiple file formats. For example we can save the DataFrame from the previous example
to a Parquet file using <code>write.df</code> (Until Spark 1.6, the default mode for writes was <code>append</code>. It was changed in Spark 1.7 to <code>error</code> to match the Scala API)</p>
<div data-lang="r">
<div class="highlight"><pre><code class="language-r" data-lang="r">write.df<span class="p">(</span>people<span class="p">,</span> path<span class="o">=</span><span class="s">&quot;people.parquet&quot;</span><span class="p">,</span> <span class="kn">source</span><span class="o">=</span><span class="s">&quot;parquet&quot;</span><span class="p">,</span> mode<span class="o">=</span><span class="s">&quot;overwrite&quot;</span><span class="p">)</span></code></pre></div>
</div>
<h3 id="from-hive-tables">From Hive tables</h3>
<p>You can also create SparkR DataFrames from Hive tables. To do this we will need to create a HiveContext which can access tables in the Hive MetaStore. Note that Spark should have been built with <a href="building-spark.html#building-with-hive-and-jdbc-support">Hive support</a> and more details on the difference between SQLContext and HiveContext can be found in the <a href="sql-programming-guide.html#starting-point-sqlcontext">SQL programming guide</a>.</p>
<div data-lang="r">
<div class="highlight"><pre><code class="language-r" data-lang="r"><span class="c1"># sc is an existing SparkContext.</span>
hiveContext <span class="o">&lt;-</span> sparkRHive.init<span class="p">(</span>sc<span class="p">)</span>
sql<span class="p">(</span>hiveContext<span class="p">,</span> <span class="s">&quot;CREATE TABLE IF NOT EXISTS src (key INT, value STRING)&quot;</span><span class="p">)</span>
sql<span class="p">(</span>hiveContext<span class="p">,</span> <span class="s">&quot;LOAD DATA LOCAL INPATH &#39;examples/src/main/resources/kv1.txt&#39; INTO TABLE src&quot;</span><span class="p">)</span>
<span class="c1"># Queries can be expressed in HiveQL.</span>
results <span class="o">&lt;-</span> sql<span class="p">(</span>hiveContext<span class="p">,</span> <span class="s">&quot;FROM src SELECT key, value&quot;</span><span class="p">)</span>
<span class="c1"># results is now a DataFrame</span>
<span class="kp">head</span><span class="p">(</span>results<span class="p">)</span>
<span class="c1">## key value</span>
<span class="c1">## 1 238 val_238</span>
<span class="c1">## 2 86 val_86</span>
<span class="c1">## 3 311 val_311</span></code></pre></div>
</div>
<h2 id="dataframe-operations">DataFrame Operations</h2>
<p>SparkR DataFrames support a number of functions to do structured data processing.
Here we include some basic examples and a complete list can be found in the <a href="api/R/index.html">API</a> docs:</p>
<h3 id="selecting-rows-columns">Selecting rows, columns</h3>
<div data-lang="r">
<div class="highlight"><pre><code class="language-r" data-lang="r"><span class="c1"># Create the DataFrame</span>
df <span class="o">&lt;-</span> createDataFrame<span class="p">(</span>sqlContext<span class="p">,</span> faithful<span class="p">)</span>
<span class="c1"># Get basic information about the DataFrame</span>
df
<span class="c1">## DataFrame[eruptions:double, waiting:double]</span>
<span class="c1"># Select only the &quot;eruptions&quot; column</span>
<span class="kp">head</span><span class="p">(</span>select<span class="p">(</span>df<span class="p">,</span> df<span class="o">$</span>eruptions<span class="p">))</span>
<span class="c1">## eruptions</span>
<span class="c1">##1 3.600</span>
<span class="c1">##2 1.800</span>
<span class="c1">##3 3.333</span>
<span class="c1"># You can also pass in column name as strings</span>
<span class="kp">head</span><span class="p">(</span>select<span class="p">(</span>df<span class="p">,</span> <span class="s">&quot;eruptions&quot;</span><span class="p">))</span>
<span class="c1"># Filter the DataFrame to only retain rows with wait times shorter than 50 mins</span>
<span class="kp">head</span><span class="p">(</span>filter<span class="p">(</span>df<span class="p">,</span> df<span class="o">$</span>waiting <span class="o">&lt;</span> <span class="m">50</span><span class="p">))</span>
<span class="c1">## eruptions waiting</span>
<span class="c1">##1 1.750 47</span>
<span class="c1">##2 1.750 47</span>
<span class="c1">##3 1.867 48</span></code></pre></div>
</div>
<h3 id="grouping-aggregation">Grouping, Aggregation</h3>
<p>SparkR data frames support a number of commonly used functions to aggregate data after grouping. For example we can compute a histogram of the <code>waiting</code> time in the <code>faithful</code> dataset as shown below</p>
<div data-lang="r">
<div class="highlight"><pre><code class="language-r" data-lang="r"><span class="c1"># We use the `n` operator to count the number of times each waiting time appears</span>
<span class="kp">head</span><span class="p">(</span>summarize<span class="p">(</span>groupBy<span class="p">(</span>df<span class="p">,</span> df<span class="o">$</span>waiting<span class="p">),</span> count <span class="o">=</span> n<span class="p">(</span>df<span class="o">$</span>waiting<span class="p">)))</span>
<span class="c1">## waiting count</span>
<span class="c1">##1 81 13</span>
<span class="c1">##2 60 6</span>
<span class="c1">##3 68 1</span>
<span class="c1"># We can also sort the output from the aggregation to get the most common waiting times</span>
waiting_counts <span class="o">&lt;-</span> summarize<span class="p">(</span>groupBy<span class="p">(</span>df<span class="p">,</span> df<span class="o">$</span>waiting<span class="p">),</span> count <span class="o">=</span> n<span class="p">(</span>df<span class="o">$</span>waiting<span class="p">))</span>
<span class="kp">head</span><span class="p">(</span>arrange<span class="p">(</span>waiting_counts<span class="p">,</span> desc<span class="p">(</span>waiting_counts<span class="o">$</span>count<span class="p">)))</span>
<span class="c1">## waiting count</span>
<span class="c1">##1 78 15</span>
<span class="c1">##2 83 14</span>
<span class="c1">##3 81 13</span></code></pre></div>
</div>
<h3 id="operating-on-columns">Operating on Columns</h3>
<p>SparkR also provides a number of functions that can directly applied to columns for data processing and during aggregation. The example below shows the use of basic arithmetic functions.</p>
<div data-lang="r">
<div class="highlight"><pre><code class="language-r" data-lang="r"><span class="c1"># Convert waiting time from hours to seconds.</span>
<span class="c1"># Note that we can assign this to a new column in the same DataFrame</span>
df<span class="o">$</span>waiting_secs <span class="o">&lt;-</span> df<span class="o">$</span>waiting <span class="o">*</span> <span class="m">60</span>
<span class="kp">head</span><span class="p">(</span>df<span class="p">)</span>
<span class="c1">## eruptions waiting waiting_secs</span>
<span class="c1">##1 3.600 79 4740</span>
<span class="c1">##2 1.800 54 3240</span>
<span class="c1">##3 3.333 74 4440</span></code></pre></div>
</div>
<h2 id="running-sql-queries-from-sparkr">Running SQL Queries from SparkR</h2>
<p>A SparkR DataFrame can also be registered as a temporary table in Spark SQL and registering a DataFrame as a table allows you to run SQL queries over its data.
The <code>sql</code> function enables applications to run SQL queries programmatically and returns the result as a <code>DataFrame</code>.</p>
<div data-lang="r">
<div class="highlight"><pre><code class="language-r" data-lang="r"><span class="c1"># Load a JSON file</span>
people <span class="o">&lt;-</span> read.df<span class="p">(</span>sqlContext<span class="p">,</span> <span class="s">&quot;./examples/src/main/resources/people.json&quot;</span><span class="p">,</span> <span class="s">&quot;json&quot;</span><span class="p">)</span>
<span class="c1"># Register this DataFrame as a table.</span>
registerTempTable<span class="p">(</span>people<span class="p">,</span> <span class="s">&quot;people&quot;</span><span class="p">)</span>
<span class="c1"># SQL statements can be run by using the sql method</span>
teenagers <span class="o">&lt;-</span> sql<span class="p">(</span>sqlContext<span class="p">,</span> <span class="s">&quot;SELECT name FROM people WHERE age &gt;= 13 AND age &lt;= 19&quot;</span><span class="p">)</span>
<span class="kp">head</span><span class="p">(</span>teenagers<span class="p">)</span>
<span class="c1">## name</span>
<span class="c1">##1 Justin</span></code></pre></div>
</div>
<h1 id="machine-learning">Machine Learning</h1>
<p>SparkR allows the fitting of generalized linear models over DataFrames using the <a href="api/R/glm.html">glm()</a> function. Under the hood, SparkR uses MLlib to train a model of the specified family. Currently the gaussian and binomial families are supported. We support a subset of the available R formula operators for model fitting, including &#8216;~&#8217;, &#8216;.&#8217;, &#8216;:&#8217;, &#8216;+&#8217;, and &#8216;-&#8216;.</p>
<p>The <a href="api/R/summary.html">summary()</a> function gives the summary of a model produced by <a href="api/R/glm.html">glm()</a>.</p>
<ul>
<li>For gaussian GLM model, it returns a list with &#8216;devianceResiduals&#8217; and &#8216;coefficients&#8217; components. The &#8216;devianceResiduals&#8217; gives the min/max deviance residuals of the estimation; the &#8216;coefficients&#8217; gives the estimated coefficients and their estimated standard errors, t values and p-values. (It only available when model fitted by normal solver.)</li>
<li>For binomial GLM model, it returns a list with &#8216;coefficients&#8217; component which gives the estimated coefficients.</li>
</ul>
<p>The examples below show the use of building gaussian GLM model and binomial GLM model using SparkR.</p>
<h2 id="gaussian-glm-model">Gaussian GLM model</h2>
<div data-lang="r">
<div class="highlight"><pre><code class="language-r" data-lang="r"><span class="c1"># Create the DataFrame</span>
df <span class="o">&lt;-</span> createDataFrame<span class="p">(</span>sqlContext<span class="p">,</span> iris<span class="p">)</span>
<span class="c1"># Fit a gaussian GLM model over the dataset.</span>
model <span class="o">&lt;-</span> glm<span class="p">(</span>Sepal_Length <span class="o">~</span> Sepal_Width <span class="o">+</span> Species<span class="p">,</span> data <span class="o">=</span> df<span class="p">,</span> family <span class="o">=</span> <span class="s">&quot;gaussian&quot;</span><span class="p">)</span>
<span class="c1"># Model summary are returned in a similar format to R&#39;s native glm().</span>
<span class="kp">summary</span><span class="p">(</span>model<span class="p">)</span>
<span class="c1">##$devianceResiduals</span>
<span class="c1">## Min Max </span>
<span class="c1">## -1.307112 1.412532</span>
<span class="c1">##</span>
<span class="c1">##$coefficients</span>
<span class="c1">## Estimate Std. Error t value Pr(&gt;|t|) </span>
<span class="c1">##(Intercept) 2.251393 0.3697543 6.08889 9.568102e-09</span>
<span class="c1">##Sepal_Width 0.8035609 0.106339 7.556598 4.187317e-12</span>
<span class="c1">##Species_versicolor 1.458743 0.1121079 13.01195 0 </span>
<span class="c1">##Species_virginica 1.946817 0.100015 19.46525 0 </span>
<span class="c1"># Make predictions based on the model.</span>
predictions <span class="o">&lt;-</span> predict<span class="p">(</span>model<span class="p">,</span> newData <span class="o">=</span> df<span class="p">)</span>
<span class="kp">head</span><span class="p">(</span>select<span class="p">(</span>predictions<span class="p">,</span> <span class="s">&quot;Sepal_Length&quot;</span><span class="p">,</span> <span class="s">&quot;prediction&quot;</span><span class="p">))</span>
<span class="c1">## Sepal_Length prediction</span>
<span class="c1">##1 5.1 5.063856</span>
<span class="c1">##2 4.9 4.662076</span>
<span class="c1">##3 4.7 4.822788</span>
<span class="c1">##4 4.6 4.742432</span>
<span class="c1">##5 5.0 5.144212</span>
<span class="c1">##6 5.4 5.385281</span></code></pre></div>
</div>
<h2 id="binomial-glm-model">Binomial GLM model</h2>
<div data-lang="r">
<div class="highlight"><pre><code class="language-r" data-lang="r"><span class="c1"># Create the DataFrame</span>
df <span class="o">&lt;-</span> createDataFrame<span class="p">(</span>sqlContext<span class="p">,</span> iris<span class="p">)</span>
training <span class="o">&lt;-</span> filter<span class="p">(</span>df<span class="p">,</span> df<span class="o">$</span>Species <span class="o">!=</span> <span class="s">&quot;setosa&quot;</span><span class="p">)</span>
<span class="c1"># Fit a binomial GLM model over the dataset.</span>
model <span class="o">&lt;-</span> glm<span class="p">(</span>Species <span class="o">~</span> Sepal_Length <span class="o">+</span> Sepal_Width<span class="p">,</span> data <span class="o">=</span> training<span class="p">,</span> family <span class="o">=</span> <span class="s">&quot;binomial&quot;</span><span class="p">)</span>
<span class="c1"># Model coefficients are returned in a similar format to R&#39;s native glm().</span>
<span class="kp">summary</span><span class="p">(</span>model<span class="p">)</span>
<span class="c1">##$coefficients</span>
<span class="c1">## Estimate</span>
<span class="c1">##(Intercept) -13.046005</span>
<span class="c1">##Sepal_Length 1.902373</span>
<span class="c1">##Sepal_Width 0.404655</span></code></pre></div>
</div>
<h1 id="r-function-name-conflicts">R Function Name Conflicts</h1>
<p>When loading and attaching a new package in R, it is possible to have a name <a href="https://stat.ethz.ch/R-manual/R-devel/library/base/html/library.html">conflict</a>, where a
function is masking another function.</p>
<p>The following functions are masked by the SparkR package:</p>
<table class="table">
<tr><th>Masked function</th><th>How to Access</th></tr>
<tr>
<td><code>cov</code> in <code>package:stats</code></td>
<td><code><pre>stats::cov(x, y = NULL, use = "everything",
method = c("pearson", "kendall", "spearman"))</pre></code></td>
</tr>
<tr>
<td><code>filter</code> in <code>package:stats</code></td>
<td><code><pre>stats::filter(x, filter, method = c("convolution", "recursive"),
sides = 2, circular = FALSE, init)</pre></code></td>
</tr>
<tr>
<td><code>sample</code> in <code>package:base</code></td>
<td><code>base::sample(x, size, replace = FALSE, prob = NULL)</code></td>
</tr>
<tr>
<td><code>table</code> in <code>package:base</code></td>
<td><code><pre>base::table(...,
exclude = if (useNA == "no") c(NA, NaN),
useNA = c("no", "ifany", "always"),
dnn = list.names(...), deparse.level = 1)</pre></code></td>
</tr>
</table>
<p>Since part of SparkR is modeled on the <code>dplyr</code> package, certain functions in SparkR share the same names with those in <code>dplyr</code>. Depending on the load order of the two packages, some functions from the package loaded first are masked by those in the package loaded after. In such case, prefix such calls with the package name, for instance, <code>SparkR::cume_dist(x)</code> or <code>dplyr::cume_dist(x)</code>.</p>
<p>You can inspect the search path in R with <a href="https://stat.ethz.ch/R-manual/R-devel/library/base/html/search.html"><code>search()</code></a></p>
<h1 id="migration-guide">Migration Guide</h1>
<h2 id="upgrading-from-sparkr-15x-to-16">Upgrading From SparkR 1.5.x to 1.6</h2>
<ul>
<li>Before Spark 1.6, the default mode for writes was <code>append</code>. It was changed in Spark 1.6.0 to <code>error</code> to match the Scala API.</li>
</ul>
</div>
<!-- /container -->
</div>
<script src="js/vendor/jquery-1.8.0.min.js"></script>
<script src="js/vendor/bootstrap.min.js"></script>
<script src="js/vendor/anchor.min.js"></script>
<script src="js/main.js"></script>
<!-- MathJax Section -->
<script type="text/x-mathjax-config">
MathJax.Hub.Config({
TeX: { equationNumbers: { autoNumber: "AMS" } }
});
</script>
<script>
// Note that we load MathJax this way to work with local file (file://), HTTP and HTTPS.
// We could use "//cdn.mathjax...", but that won't support "file://".
(function(d, script) {
script = d.createElement('script');
script.type = 'text/javascript';
script.async = true;
script.onload = function(){
MathJax.Hub.Config({
tex2jax: {
inlineMath: [ ["$", "$"], ["\\\\(","\\\\)"] ],
displayMath: [ ["$$","$$"], ["\\[", "\\]"] ],
processEscapes: true,
skipTags: ['script', 'noscript', 'style', 'textarea', 'pre']
}
});
};
script.src = ('https:' == document.location.protocol ? 'https://' : 'http://') +
'cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML';
d.getElementsByTagName('head')[0].appendChild(script);
}(document));
</script>
</body>
</html>