Added Spark accumulators snippets.

spark/spark.ipynb

@@ -1,7 +1,7 @@
 {
  "metadata": {
   "name": "",
-  "signature": "sha256:cf2651f340403267ecf82871c01a4f9deb80df25b8345be4f56dc02db970a5a7"
+  "signature": "sha256:f13a3255902ddadfe3c0b99e722529cfe44804d81633f80f0f7681afcf3acb8a"
  },
  "nbformat": 3,
  "nbformat_minor": 0,
@@ -26,7 +26,8 @@
       "* Configuring Spark Applications\n",
       "* Streaming\n",
       "* Streaming with States\n",
-      "* Broadcast Variables"
+      "* Broadcast Variables\n",
+      "* Accumulators"
      ]
     },
     {
@@ -1143,6 +1144,82 @@
      "language": "python",
      "metadata": {},
      "outputs": []
+    },
+    {
+     "cell_type": "markdown",
+     "metadata": {},
+     "source": [
+      "## Accumulators"
+     ]
+    },
+    {
+     "cell_type": "markdown",
+     "metadata": {},
+     "source": [
+      "Create an accumulator:"
+     ]
+    },
+    {
+     "cell_type": "code",
+     "collapsed": false,
+     "input": [
+      "txt_count = sc.accumulator(0)"
+     ],
+     "language": "python",
+     "metadata": {},
+     "outputs": []
+    },
+    {
+     "cell_type": "markdown",
+     "metadata": {},
+     "source": [
+      "Count the number of txt files in the RDD:"
+     ]
+    },
+    {
+     "cell_type": "code",
+     "collapsed": false,
+     "input": [
+      "my_data = sc.textFile(filePath)\n",
+      "my_data.foreach(lambda line: if '.txt' in line: txt_count.add(1))"
+     ],
+     "language": "python",
+     "metadata": {},
+     "outputs": []
+    },
+    {
+     "cell_type": "markdown",
+     "metadata": {},
+     "source": [
+      "Count the number of file types encountered:"
+     ]
+    },
+    {
+     "cell_type": "code",
+     "collapsed": false,
+     "input": [
+      "jpg_count = sc.accumulator(0)\n",
+      "html_count = sc.accumulator(0)\n",
+      "css_count = sc.accumulator(0)\n",
+      "\n",
+      "def countFileType(s):\n",
+      "    if '.jpg' in s: jpg_count.add(1)\n",
+      "    elif '.html' in s: html_count.add(1)\n",
+      "    elif '.css' in s: css_count.add(1)\n",
+      "\n",
+      "filename=\"hdfs://logs/*\"\n",
+      "\n",
+      "logs = sc.textFile(filename)\n",
+      "logs.foreach(lambda line: countFileType(line))\n",
+      "\n",
+      "print  'File Type Totals:'\n",
+      "print '.css files: ', css_count.value\n",
+      "print '.html files: ', html_count.value\n",
+      "print '.jpg files: ', jpg_count.value"
+     ],
+     "language": "python",
+     "metadata": {},
+     "outputs": []
     }
    ],
    "metadata": {}