{"id":263,"date":"2020-04-29T19:26:11","date_gmt":"2020-04-29T10:26:11","guid":{"rendered":"http:\/\/localhost:8000\/?p=263"},"modified":"2021-01-16T19:29:41","modified_gmt":"2021-01-16T10:29:41","slug":"pyspark-samples","status":"publish","type":"post","link":"http:\/\/localhost:8000\/2020\/04\/pyspark-samples.html","title":{"rendered":"PySpark\u306e\u5b9f\u88c5\u30b5\u30f3\u30d7\u30eb\u3068\u5b9f\u884c\u65b9\u6cd5"},"content":{"rendered":"<p>\u6700\u8fd1Spark\u3092\u89e6\u308b\u6a5f\u4f1a\u304c\u3042\u3063\u3066\u3001\u5c11\u3057\u3060\u3051\u52c9\u5f37\u3057\u305f\u306e\u3067\u30e1\u30e2\u304c\u3066\u3089\u6b8b\u3057\u3066\u304a\u304d\u307e\u3059\u3002<\/p>\n<h2>Spark\u306e\u5206\u6563\u51e6\u7406\u306e\u4ed5\u7d44\u307f<\/h2>\n<h3>Spark\u3068\u306f<\/h3>\n<ul>\n<li>\u9ad8\u901f\u3067\u6c4e\u7528\u7684\u306a\u5206\u6563\u51e6\u7406\u30b7\u30b9\u30c6\u30e0<\/li>\n<li>\u5206\u6563\u30c7\u30fc\u30bf\uff08RDD\uff09\u3092DISK\u3092\u4ecb\u3055\u305a\u306b\u30e1\u30e2\u30ea\u4e0a\u306b\u6301\u3064\u306e\u3067\u3001Hadoop\u306e100\u500d\u3050\u3089\u3044\u9ad8\u901f<\/li>\n<li>Java, Scala, Python, R\u306a\u3069\u306eAPI\u3092\u63d0\u4f9b<\/li>\n<li>Spark SQL, MLlib, GraphX, Spark Streaming\u306a\u3069\u306e\u30ea\u30c3\u30c1\u306a\u30c4\u30fc\u30eb\u3092\u63d0\u4f9b<\/li>\n<\/ul>\n<h3>\u5206\u6563\u51e6\u7406\u30b7\u30b9\u30c6\u30e0\u306e\u69cb\u6210\u8981\u7d20<\/h3>\n<p>\u753b\u50cf\u306f\u3001<a href=\"https:\/\/spark.apache.org\/docs\/latest\/cluster-overview.html\">\u3053\u3061\u3089<\/a>\u304b\u3089\u304a\u501f\u308a\u3057\u307e\u3057\u305f\u3002<br \/>\n<img src=\"http:\/\/localhost:8000\/wp-content\/uploads\/2020\/04\/20200429115647-400x179.png\" alt=\"\" \/><\/p>\n<ul>\n<li>Driver Program\n<ul>\n<li>Master Node\u3067\u5b9f\u884c\u3055\u308c\u308b\u8d77\u70b9\u3068\u306a\u308b\u30d7\u30ed\u30b0\u30e9\u30e0<\/li>\n<li>SparkContext\u3092\u4f5c\u6210\u3057\u3001RDD\u3092\u751f\u6210\u3057\u3066\u3001Task\u3092\u5b9f\u884c\u3057\u3066\u3044\u304f<\/li>\n<\/ul>\n<\/li>\n<li>SparkContext\n<ul>\n<li>Spark\u306e\u8272\u3005\u306a\u6a5f\u80fd\u3078\u306e\u30a8\u30f3\u30c8\u30ea\u30fc\u30dd\u30a4\u30f3\u30c8<\/li>\n<li>ClusterManager\u3092\u901a\u3058\u3066\u30af\u30e9\u30b9\u30bf\u30fc\u3092\u64cd\u4f5c\u3059\u308b<\/li>\n<li>DAG Scheduler: \u4e00\u5ea6\u306b\u4e26\u5217\u51e6\u7406\u53ef\u80fd\u306a\u7bc4\u56f2\u3092Stage\u306b\u5206\u5272\u3057\u3066\u3001 Stage\u5185\u306f\u8907\u6570\u306eTask\u3067\u69cb\u6210\u3059\u308b\u3002\u305d\u308c\u3092DAG\uff08\u6709\u5411\u975e\u5de1\u56de\u30b0\u30e9\u30d5\uff09\u306e\u5f62\u5f0f\u3067\u8868\u73fe\u3059\u308b\u306e\u3067DAG Scheduler\u3068\u3044\u3046\u3089\u3057\u3044\u3002<\/li>\n<li>Task Scheduler: DAG Scheduler\u3067\u30b9\u30b1\u30b8\u30e5\u30fc\u30eb\u3055\u308c\u305f\u9806\u756a\u306bTask\u3092Cluster Manager\u306b\u30ed\u30f3\u30c1\u3059\u308b<\/li>\n<li><a href=\"https:\/\/data-flair.training\/blogs\/dag-in-apache-spark\/\">\u3053\u3061\u3089<\/a>\u306e\u30b5\u30a4\u30c8\u304c\u308f\u304b\u308a\u3084\u3059\u304b\u3063\u305f<\/li>\n<\/ul>\n<\/li>\n<li>ClusterManager\n<ul>\n<li>\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u3092\u5b9f\u884c\u3059\u308b\u305f\u3081\u306b\u5fc5\u8981\u306a\u30ea\u30bd\u30fc\u30b9\u306e\u5272\u308a\u5f53\u3066\u6e21\u3063\u3066\u304d\u305fTask\u3092Executer\u306b\u6e21\u3059<\/li>\n<li>Spark Standalone Cluster, Hadoop YARN, Apache Mesos\u306a\u3069<\/li>\n<\/ul>\n<\/li>\n<li>Executor\n<ul>\n<li>Worker Node\u4e0a\u3067JVM\u30b3\u30f3\u30c6\u30ca\u8d77\u52d5\u3055\u308c\u308b\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u5c02\u7528\u306e\u30d7\u30ed\u30bb\u30b9<\/li>\n<li>Task\u3092\u5b9f\u884c\u3059\u308b\u3002\u307e\u305fCache Data\u3092\u4fdd\u6301\u3059\u308b<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h3>\u51e6\u7406\u306e\u6d41\u308c<\/h3>\n<p>\u81ea\u5206\u306e\u7406\u89e3\u3067\u306f\u3001\u4ee5\u4e0b\u306e\u3088\u3046\u306a\u51e6\u7406\u306e\u6d41\u308c\u306b\u306a\u3063\u3066\u3044\u308b\u307f\u305f\u3044\u3067\u3059\u3002<br \/>\n<img src=\"http:\/\/localhost:8000\/wp-content\/uploads\/2020\/04\/20200429121743-600x291.png\" alt=\"\" \/><\/p>\n<h2>\u5206\u6563\u51e6\u7406\u306e\u5b9f\u88c5\u30b5\u30f3\u30d7\u30eb<\/h2>\n<h3>\u51e6\u7406\u5185\u5bb9<\/h3>\n<p>\u8a66\u9a13\u306e\u7d50\u679c\u304b\u3089\u300c\u5404\u6559\u79d1\u306e\u5e73\u5747\u70b9\u300d\u3068\u300c4\u6559\u79d1\u5408\u8a08\u306e\u5e73\u5747\u70b9\u300d\u3092\u7b97\u51fa\u3059\u308b\u3068\u3044\u3046\u6bd4\u8f03\u7684\u30b7\u30f3\u30d7\u30eb\u306a\u51e6\u7406\u3092\u8003\u3048\u3066\u307f\u3088\u3046\u3068\u601d\u3044\u307e\u3059\u3002<\/p>\n<h3>\u30bd\u30fc\u30b9\u30b3\u30fc\u30c9<\/h3>\n<pre><code class=\"language-python\">from pyspark import SparkContext, RDD\nfrom typing import List, Dict\n# input\u30c7\u30fc\u30bf\uff08\u8a66\u9a13\u306e\u7d50\u679c\uff09\ninput_data: List[Dict[str, int]] = [\n   {&#039;\u56fd\u8a9e&#039;: 86, &#039;\u7b97\u6570&#039;: 57, &#039;\u7406\u79d1&#039;: 45, &#039;\u793e\u4f1a&#039;: 100},\n   {&#039;\u56fd\u8a9e&#039;: 67, &#039;\u7b97\u6570&#039;: 12, &#039;\u7406\u79d1&#039;: 43, &#039;\u793e\u4f1a&#039;: 54},\n   {&#039;\u56fd\u8a9e&#039;: 98, &#039;\u7b97\u6570&#039;: 98, &#039;\u7406\u79d1&#039;: 78, &#039;\u793e\u4f1a&#039;: 69},\n]\n# SparkContext, RDD\u4f5c\u6210\nsc: SparkContext = SparkContext(appName=&#039;spark_sample&#039;)\nrdd: RDD = sc.parallelize(input_data)\n# \u5404\u6559\u79d1\u304a\u3088\u3073\u5408\u8a08\u70b9\u306e\u5e73\u5747\u70b9\u3092\u8a08\u7b97\noutput: Dict[str, float] = rdd\\\n   .map(lambda x: x.update(\u5408\u8a08=sum(x.values())) or x)\\\n   .flatMap(lambda x: x.items())\\\n   .groupByKey()\\\n   .map(lambda x: (x[0], round(sum(x[1]) \/ len(x[1]), 2)))\\\n   .collect()<\/code><\/pre>\n<h3>\u89e3\u8aac<\/h3>\n<p>\u307e\u305a\u3001SparkContext\u3092\u751f\u6210\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<pre><code class=\"language-pythonn\">sc: SparkContext = SparkContext(appName=&#039;spark_sample&#039;)<\/code><\/pre>\n<p>\u524d\u8ff0\u306e\u69cb\u6210\u8981\u7d20\u306e\u3068\u3053\u308d\u3067\u8aac\u660e\u3057\u305f\u3084\u3064\u3067\u3059\u3002DAG Scheduler\u3084Task Scheduler\u306e\u5f79\u5272\u3092\u679c\u305f\u3057\u3066\u304f\u308c\u307e\u3059\u3002<\/p>\n<p>\u6b21\u306b\u3001RDD\u3092\u751f\u6210\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<pre><code class=\"language-python\">rdd: RDD = sc.parallelize(input_data)<\/code><\/pre>\n<p>RDD\u306f\u5fa9\u5143\u529b\u306e\u3042\u308b\u5206\u6563\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3067\u3001\u305f\u3060List\u3092\u5206\u5272\u3057\u305f\u3060\u3051\u3067\u306a\u304f\u3001\u3044\u308d\u3093\u306a\u7279\u5fb4\u304c\u3042\u308a\u307e\u3059\u3002<\/p>\n<ul>\n<li>Fault Tolerance: \u5931\u6557\u3057\u3066\u3082\u3044\u3044\u611f\u3058\u306b\u5fa9\u65e7\u3057\u3066\u304f\u308c\u308b<\/li>\n<li>Immutable: \u5909\u63db\u51e6\u7406\u3092\u884c\u3046\u5834\u5408\u306f\u3001\u5fc5\u305a\u65b0\u3057\u3044RDD\u3092\u4f5c\u6210\u3059\u308b\u3002\u5143\u30c7\u30fc\u30bf\u304c\u4fdd\u6301\u3055\u308c\u3066\u3044\u308b\u306e\u3067\u30ea\u30c8\u30e9\u30a4\u304c\u3067\u304d\u308b<\/li>\n<li>Lazy Evaluation: <code>collect<\/code>\u3084<code>foreach<\/code>\u306a\u3069\u306eAction\u3068\u547c\u3070\u308c\u308bTask\u304c\u5b9f\u884c\u3055\u308c\u308b\u307e\u3067\u51e6\u7406\u3055\u308c\u306a\u3044\u3002Action\u304c\u5b9f\u884c\u3055\u308c\u305f\u30bf\u30a4\u30df\u30f3\u30b0\u3067\u5fc5\u8981\u306aTask\u3060\u3051\u5b9f\u884c\u3055\u308c\u308b\uff08=job\u304csubmit\u3055\u308c\u308b\uff09<\/li>\n<\/ul>\n<p>\u306a\u3069\u306e\u7279\u5fb4\u3092\u6301\u3063\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\u6700\u5f8c\u306b\u3001RDD\u306b\u5bfe\u3057\u3066\u4e00\u9023\u306e\u5909\u63db\u30bf\u30b9\u30af\u884c\u3044\u3001\u7d50\u679c\u3092\u53d6\u5f97\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<pre><code class=\"language-python\">output: Dict[str, float] = rdd\\\n   .map(lambda x: x.update(\u5408\u8a08=sum(x.values())) or x)\\\n   .flatMap(lambda x: x.items())\\\n   .groupByKey()\\\n   .map(lambda x: (x[0], round(sum(x[1]) \/ len(x[1]), 2)))\\\n   .collect()<\/code><\/pre>\n<p>\u5177\u4f53\u7684\u306b\u306f\u3001\u4ee5\u4e0b\u306e\u3088\u3046\u306a\u6d41\u308c\u3067\u51e6\u7406\u3092\u884c\u306a\u3063\u3066\u3044\u307e\u3059\u3002<\/p>\n<ol>\n<li>\u5408\u8a08\u3092\u5404record\u306b\u8ffd\u52a0  <code>{ \u2018\u56fd\u8a9e\u2019: 86, \u2018\u7b97\u6570\u2019: 57, \u2026, \u2018\u5408\u8a08\u2019: 288 }<\/code><\/li>\n<li>List[Tuple[\u6559\u79d1, \u70b9\u6570]]\u306b\u5909\u63db\uff08\u7e26\u6301\u3061\u306b\uff09 <code>[(\u2018\u56fd\u8a9e\u2019, 86), (\u2018\u7b97\u6570\u2019, 57), \u2026 , (\u2018\u5408\u8a08\u2019, 288)]<\/code><\/li>\n<li>\u6559\u79d1\u3067\u30b0\u30eb\u30fc\u30d4\u30f3\u30b0 <code>(\u2018\u56fd\u8a9e\u2019, [86, 67, 98])<\/code><\/li>\n<li>\u5e73\u5747\u70b9\u3092\u7b97\u51fa <code>(\u2018\u56fd\u8a9e\u2019, 83.67)<\/code><\/li>\n<li>\u5168record\u3092\u96c6\u7d04 <code>[(\u2018\u56fd\u8a9e\u2019, 83.67), (\u2018\u7b97\u6570\u2019, 55.67), \u2026, (\u2018\u5408\u8a08\u2019, 269.0)]<\/code><\/li>\n<\/ol>\n<p>Stage\u69cb\u6210\u306f\u3053\u3093\u306a\u611f\u3058\u3067\u3059\u3002<\/p>\n<p>[f:id:rinoguchi:20200429125214p:plain:w500]<\/p>\n<p><code>collect<\/code>\u306a\u3069\u306eAction Task\u306e\u4ed6\u306b\u3001<code>groupByKey<\/code>\u306e\u3088\u3046\u306aShuffle Task\u3067\u3082Stage\u304c\u5206\u5272\u3055\u308c\u307e\u3059\u3002<code>ByKey<\/code>\u3067\u7d42\u308f\u308b\u3088\u3046\u306aTask\u304cShuffle Task\u3067\u3059\u3002\u4e00\u3064\u306eStage\u306e\u4e00\u9023\u306eTask\u306f\u4e00\u3064\u306eNode\u3067\u4e00\u62ec\u3067\u51e6\u7406\u3055\u308c\u307e\u3059\u3002<\/p>\n<h2>\u5206\u6563\u51e6\u7406\u306e\u5b9f\u884c<\/h2>\n<p>\u3053\u308c\u304b\u30893\u3064\u306e\u65b9\u6cd5\u3067\u5148\u307b\u3069\u5b9f\u88c5\u3057\u305f\u5206\u6563\u51e6\u7406\u30d7\u30ed\u30b0\u30e9\u30e0\uff08main.py\uff09\u3092\u5b9f\u884c\u3057\u3066\u307f\u305f\u3044\u3068\u3082\u601d\u3044\u307e\u3059\u3002\u30d7\u30ed\u30b0\u30e9\u30e0\u306f\u3069\u306e\u65b9\u6cd5\u3067\u3082\u5909\u66f4\u306e\u5fc5\u8981\u306f\u3042\u308a\u307e\u305b\u3093\u3002<\/p>\n<h3>pyspark\u30e9\u30a4\u30d6\u30e9\u30ea\u3092\u4f7f\u3063\u3066\u5b9f\u884c<\/h3>\n<h4>\u5b9f\u884c\u65b9\u6cd5<\/h4>\n<pre><code class=\"language-sh\"># \u5fc5\u8981\u306a\u30e9\u30a4\u30d6\u30e9\u30ea\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\npip install pyspark\npip install py4j\n\n# \u5b9f\u884c\npython main.py<\/code><\/pre>\n<h4>\u7279\u5fb4<\/h4>\n<ul>\n<li>\u901a\u5e38\u306epython\u30a2\u30d7\u30ea\u3068\u5168\u304f\u540c\u69d8\u306b\u5b9f\u884c\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u3066\u304a\u624b\u8efd<\/li>\n<li>\u30af\u30e9\u30b9\u30bf\u6e96\u5099\u4e0d\u8981<\/li>\n<li>\u30a2\u30d7\u30ea\u306e\u52d5\u4f5c\u78ba\u8a8d\u306f\u307b\u3068\u3093\u3069\u3053\u308c\u3067\u3044\u3044\u3068\u601d\u3046<\/li>\n<li>\u5b9f\u904b\u7528\u3067\u306f\u4f7f\u3048\u306a\u3044<\/li>\n<\/ul>\n<h3>Spark Standalone Mode\u3067\u5b9f\u884c<\/h3>\n<h4>\u5b9f\u884c\u65b9\u6cd5<\/h4>\n<pre><code class=\"language-sh\"># apache-spark\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\nbrew install apache-spark\nexport SPARK_HOME=\/usr\/local\/Cellar\/apache-spark\/2.4.5\/libexec\n\n# Master Node\u3092\u7acb\u3061\u4e0a\u3052\u308b\n# \u7acb\u3061\u4e0a\u3052\u305f\u5f8c\u306b\u3001http:\/\/localhost:8080\u306b\u30a2\u30af\u30bb\u30b9\u3057\u3066\u3001Master-URL\u3092\u78ba\u8a8d\u3059\u308b\n# \u4f8b) spark:\/\/hogeMacBook-Pro:7077\n${SPARK_HOME}\/sbin\/start-master.sh\n\n# Worker Node\u3092\u7acb\u3061\u4e0a\u3052\u308b\n${SPARK_HOME}\/sbin\/start-slave.sh -c 2 spark:\/\/InoguchinoMacBook-Pro:7077\n\n# job\u3092submit\u3059\u308b\nspark-submit main.py --master spark:\/\/InoguchinoMacBook-Pro:7077<\/code><\/pre>\n<h4>\u7279\u5fb4<\/h4>\n<ul>\n<li>\u30ed\u30fc\u30ab\u30eb\u30de\u30b7\u30f3\u4e0a\u306b\u30af\u30e9\u30b9\u30bf\uff08master &amp; workers\uff09\u3092\u69cb\u7bc9\u3057\u3066\u3001\u30af\u30e9\u30b9\u30bf\u4e0a\u3067\u30b8\u30e7\u30d6\u3092\u5b9f\u884c\u3055\u305b\u308b\u305f\u3081\u3001\u30af\u30e9\u30b9\u30bf\u3092\u6e96\u5099\u3059\u308b\u5fc5\u8981\u304c\u3042\u308b\n<ul>\n<li>\u3044\u304f\u3064\u304b\u306e\u30b3\u30de\u30f3\u30c9\u3092\u6253\u3064\u5fc5\u8981\u304c\u3042\u308a\u591a\u5c11\u624b\u9593<\/li>\n<\/ul>\n<\/li>\n<li>SparkUI\u3067\u3001Event Timeline\u3092\u78ba\u8a8d\u3057\u305f\u308a\u3001DAG\uff08\u30bf\u30b9\u30af\u306e\u6d41\u308c\u3092\u8868\u3057\u305f\u6709\u5411\u975e\u5faa\u74b0\u30b0\u30e9\u30d5\uff09\u3092\u78ba\u8a8d\u3067\u304d\u308b\u306e\u3067\u3001\u6027\u80fd\u6539\u5584\u306e\u7528\u9014\u3067\u4f7f\u3046<\/li>\n<li>\u5b9f\u904b\u7528\u3067\u306f\u4f7f\u3046\u610f\u5473\u306f\u306a\u3055\u305d\u3046\n<ul>\n<li>\u540c\u4e00\u30de\u30b7\u30f3\u5185\u3067\u30de\u30eb\u30c1\u30b3\u30a2\u3067\u4e26\u5217\u51e6\u7406\u3057\u305f\u3044\u3060\u3051\u306a\u3089\u3001<code>concurrent.futures<\/code>\u3068\u304b\u3067\u3082\u3067\u304d\u308b\u306e\u3067<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h3>Google Cloud Dataproc\u3067\u5b9f\u884c<\/h3>\n<h4>Dataproc\u306e\u74b0\u5883\u6e96\u5099<\/h4>\n<p><a href=\"http:\/\/lab.astamuse.co.jp\/entry\/2020\/04\/08\/113000\">\u3053\u3061\u3089<\/a>\u306e\u300cDataproc\u5468\u308a\u306e\u74b0\u5883\u6e96\u5099\u300d\u306e\u9805\u3092\u53c2\u7167\u3057\u3001Dataproc\u3092\u6709\u52b9\u5316\u3057\u3066\u30b5\u30fc\u30d3\u30b9\u30a2\u30ab\u30a6\u30f3\u30c8\u3092\u4f5c\u6210\u3057\u5fc5\u8981\u306a\u6a29\u9650\u3092\u4e0e\u3048\u307e\u3059\u3002<\/p>\n<h4>\u5b9f\u884c\u65b9\u6cd5<\/h4>\n<pre><code class=\"language-sh\"># \u30b5\u30fc\u30d3\u30b9\u30a2\u30ab\u30a6\u30f3\u30c8\u3092\u6709\u52b9\u5316\ngcloud auth activate-service-account --key-file {key_file_json_path} --project {project_name}\n\n# \u30af\u30e9\u30b9\u30bf\u7acb\u3061\u4e0a\u3052\ngcloud dataproc clusters create spark-sample-cluster\\\n  --bucket={bucket_name} --region=asia-east1 --image-version=1.4\\  \n  --master-machine-type=n1-standard-1 --worker-machine-type=n1-standard-1\\\n  --num-workers=2 --worker-boot-disk-size=256 --max-idle=1h\n\n# job\u3092submit\ngcloud dataproc jobs submit pyspark\\\n  --cluster=spark-sample-cluster --region=asia-east1\\\n  gs:\/\/hoge_dataproc-test\/dataproc\/src\/simple\/main.py<\/code><\/pre>\n<h4>\u7279\u5fb4<\/h4>\n<ul>\n<li>Dataproc\u306fGoogle\u304c\u63d0\u4f9b\u3059\u308b\u30d5\u30eb\u30de\u30cd\u30fc\u30b8\u30c9\u306a\u5206\u6563\u51e6\u7406\u30af\u30e9\u30a6\u30c9\u30b5\u30fc\u30d3\u30b9<\/li>\n<li>\u30af\u30e9\u30b9\u30bf\uff08master &amp; workers\uff09\u3092\u7d20\u4eba\u3067\u3082\u7c21\u5358\u306b\u4f5c\u308c\u308b<\/li>\n<li>\u30b3\u30f3\u30bd\u30fc\u30eb\u753b\u9762\u3001CLI\u3001\u5404\u8a00\u8a9e\u306e\u30c9\u30e9\u30a4\u30d0\u30fc\u304b\u3089\u3001\u624b\u8efd\u306b\u30af\u30e9\u30b9\u30bf\u3092\u4f5c\u6210\u3057\u305f\u308a\u3001JOB\u3092\u5b9f\u884c\u3057\u305f\u308a\u3067\u304d\u308b<\/li>\n<li>1vCPU \u00d7 1hour = 1\u30bb\u30f3\u30c8 = 0.01\u30c9\u30eb<\/li>\n<\/ul>\n<h2>\u307e\u3068\u3081<\/h2>\n<p>\u500b\u4eba\u7684\u306b\u306f\u3001<\/p>\n<ol>\n<li>\u307e\u305apyspark\u30e9\u30a4\u30d6\u30e9\u30ea\u3067\u5b9f\u88c5\u30fb\u52d5\u4f5c\u78ba\u8a8d\u306e\u30a4\u30c6\u30ec\u30fc\u30b7\u30e7\u30f3\u3092\u884c\u3044<\/li>\n<li>Spark Standalone Mode\u3067\u8efd\u304fDAG Schedule\u3084\u5404Task\u306e\u5b9f\u884c\u6642\u9593\u3092\u78ba\u8a8d\u3057\u305f\u4e0a\u3067<\/li>\n<li>\u6700\u7d42\u7684\u306bDataproc\u3067\u672c\u756a\u904b\u7528\u3059\u308b<\/li>\n<\/ol>\n<p>\u3068\u3044\u3046\u6d41\u308c\u3067\u958b\u767a\u3092\u884c\u3044\u307e\u3057\u305f\u304c\u3001\u3053\u308c\u306f\u7d50\u69cb\u3057\u3063\u304f\u308a\u304d\u307e\u3057\u305f\u3002<\/p>\n<h2>\u53c2\u8003URL<\/h2>\n<ul>\n<li>Spark\n<ul>\n<li><a href=\"https:\/\/spark.apache.org\/docs\/latest\/index.html\">Overview - Spark 3.0.1 Documentation<\/a><\/li>\n<li><a href=\"https:\/\/techvidvan.com\/tutorials\/spark-tutorial\/\">Spark Tutorial - Apache Spark Introduction for Beginners - TechVidvan<\/a><\/li>\n<\/ul>\n<\/li>\n<li>YARN\n<ul>\n<li><a href=\"https:\/\/www.ibm.com\/developerworks\/jp\/analytics\/library\/bd-yarn-intro\/index.html\">YARN \u306e\u7d39\u4ecb<\/a><\/li>\n<\/ul>\n<\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>\u6700\u8fd1Spark\u3092\u89e6\u308b\u6a5f\u4f1a\u304c\u3042\u3063\u3066\u3001\u5c11\u3057\u3060\u3051\u52c9\u5f37\u3057\u305f\u306e\u3067\u30e1\u30e2\u304c\u3066\u3089\u6b8b\u3057\u3066\u304a\u304d\u307e\u3059\u3002 Spark\u306e\u5206\u6563\u51e6\u7406\u306e\u4ed5\u7d44\u307f Spark\u3068\u306f \u9ad8\u901f\u3067\u6c4e\u7528\u7684\u306a\u5206\u6563\u51e6\u7406\u30b7\u30b9\u30c6\u30e0 \u5206\u6563\u30c7\u30fc\u30bf\uff08RDD\uff09\u3092DISK\u3092\u4ecb\u3055\u305a\u306b\u30e1\u30e2\u30ea\u4e0a\u306b\u6301\u3064\u306e\u3067\u3001Hadoop\u306e100\u500d\u3050\u3089\u3044\u9ad8\u901f Java, Scala, Python, R\u306a\u3069\u306eAPI\u3092\u63d0\u4f9b Spark SQL, MLlib, GraphX, Spark Streaming\u306a\u3069\u306e\u30ea\u30c3\u30c1\u306a\u30c4\u30fc\u30eb\u3092\u63d0\u4f9b \u5206\u6563\u51e6\u7406\u30b7\u30b9\u30c6\u30e0\u306e\u69cb\u6210\u8981\u7d20 \u753b\u50cf\u306f\u3001\u3053\u3061\u3089\u304b\u3089\u304a\u501f\u308a\u3057\u307e\u3057\u305f\u3002 Driver Program Master Node\u3067\u5b9f\u884c\u3055\u308c\u308b\u8d77\u70b9\u3068\u306a\u308b\u30d7\u30ed\u30b0\u30e9\u30e0 SparkContext\u3092\u4f5c\u6210\u3057\u3001RDD\u3092\u751f\u6210\u3057\u3066\u3001Task\u3092\u5b9f\u884c\u3057\u3066\u3044\u304f SparkContext Spark\u306e\u8272\u3005\u306a\u6a5f\u80fd\u3078\u306e\u30a8\u30f3\u30c8\u30ea\u30fc\u30dd\u30a4\u30f3\u30c8 ClusterManager\u3092\u901a\u3058\u3066\u30af\u30e9\u30b9\u30bf\u30fc\u3092\u64cd\u4f5c\u3059\u308b DAG Sche <a href=\"http:\/\/localhost:8000\/2020\/04\/pyspark-samples.html\" class=\"read-more button-fancy -red\"><span class=\"btn-arrow\"><\/span><span class=\"twp-read-more text\">Continue Reading<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[7,24],"tags":[],"_links":{"self":[{"href":"http:\/\/localhost:8000\/wp-json\/wp\/v2\/posts\/263"}],"collection":[{"href":"http:\/\/localhost:8000\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/localhost:8000\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/localhost:8000\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/localhost:8000\/wp-json\/wp\/v2\/comments?post=263"}],"version-history":[{"count":1,"href":"http:\/\/localhost:8000\/wp-json\/wp\/v2\/posts\/263\/revisions"}],"predecessor-version":[{"id":266,"href":"http:\/\/localhost:8000\/wp-json\/wp\/v2\/posts\/263\/revisions\/266"}],"wp:attachment":[{"href":"http:\/\/localhost:8000\/wp-json\/wp\/v2\/media?parent=263"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/localhost:8000\/wp-json\/wp\/v2\/categories?post=263"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/localhost:8000\/wp-json\/wp\/v2\/tags?post=263"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}