tutorials/basic_example.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Basic Example\n",
    "\n",
    "Here is a basic example of running a `VerificationSuite` with a couple `checks` and then filtering them based on their results. \n",
    "\n",
    "We'll start by creating a Spark session and a small sample dataframe."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pydeequ\n",
    "\n",
    "import sagemaker_pyspark\n",
    "from pyspark.sql import SparkSession, Row\n",
    "\n",
    "classpath = \":\".join(sagemaker_pyspark.classpath_jars()) # aws-specific jars\n",
    "\n",
    "spark = (SparkSession\n",
    "    .builder\n",
    "    .config(\"spark.driver.extraClassPath\", classpath)\n",
    "    .config(\"spark.jars.packages\", pydeequ.deequ_maven_coord)\n",
    "    .config(\"spark.jars.excludes\", pydeequ.f2j_maven_coord)\n",
    "    .getOrCreate())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = spark.sparkContext.parallelize([\n",
    "    Row(a=\"foo\", b=1, c=5),\n",
    "    Row(a=\"bar\", b=2, c=6),\n",
    "    Row(a=\"baz\", b=3, c=None)]).toDF()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, we will be importing the necessary `PyDeequ` modules for running a VerificationSuite with Checks. We will be checking the following: \n",
    "\n",
    "- does `df` have a size of at least 3? \n",
    "- does the `b` column have a minimum value of 0? \n",
    "- is the `c` column complete? \n",
    "- is the `a` column unique? \n",
    "- are the values of `a` column contained in \"foo\", \"bar\", and \"baz\"? \n",
    "- are the values in `b` colum non-negative? \n",
    "\n",
    "Once these checks are run, we'll display out the dataframe to see the results!\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Python Callback server started!\n",
      "+----------------+-----------+------------+--------------------+-----------------+--------------------+\n",
      "|           check|check_level|check_status|          constraint|constraint_status|  constraint_message|\n",
      "+----------------+-----------+------------+--------------------+-----------------+--------------------+\n",
      "|Integrity checks|      Error|       Error|SizeConstraint(Si...|          Success|                    |\n",
      "|Integrity checks|      Error|       Error|MinimumConstraint...|          Failure|Value: 1.0 does n...|\n",
      "|Integrity checks|      Error|       Error|CompletenessConst...|          Failure|Value: 0.66666666...|\n",
      "|Integrity checks|      Error|       Error|UniquenessConstra...|          Success|                    |\n",
      "|Integrity checks|      Error|       Error|ComplianceConstra...|          Success|                    |\n",
      "|Integrity checks|      Error|       Error|ComplianceConstra...|          Success|                    |\n",
      "+----------------+-----------+------------+--------------------+-----------------+--------------------+\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from pydeequ.checks import *\n",
    "from pydeequ.verification import *\n",
    "\n",
    "check = Check(spark, CheckLevel.Error, \"Integrity checks\")\n",
    "\n",
    "checkResult = VerificationSuite(spark) \\\n",
    "    .onData(df) \\\n",
    "    .addCheck(\n",
    "        check.hasSize(lambda x: x >= 3) \\\n",
    "        .hasMin(\"b\", lambda x: x == 0) \\\n",
    "        .isComplete(\"c\")  \\\n",
    "        .isUnique(\"a\")  \\\n",
    "        .isContainedIn(\"a\", [\"foo\", \"bar\", \"baz\"]) \\\n",
    "        .isNonNegative(\"b\")) \\\n",
    "    .run()\n",
    "\n",
    "checkResult_df = VerificationResult.checkResultsAsDataFrame(spark, checkResult)\n",
    "checkResult_df.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It's nice to see those as a dataframe, but we noticed a couple **Failures** in the `constraint_status` column! Let's filter them by accessing the `checkResults` property of our run."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "We found errors in the data, the following constraints were not satisfied:\n",
      "\tMinimumConstraint(Minimum(b,None)) failed because: Value: 1.0 does not meet the constraint requirement!\n",
      "\tCompletenessConstraint(Completeness(c,None)) failed because: Value: 0.6666666666666666 does not meet the constraint requirement!\n"
     ]
    }
   ],
   "source": [
    "if checkResult.status == \"Success\": \n",
    "    print('The data passed the test, everything is fine!')\n",
    "\n",
    "else:\n",
    "    print('We found errors in the data, the following constraints were not satisfied:')\n",
    "    \n",
    "    for check_json in checkResult.checkResults:\n",
    "        if check_json['constraint_status'] != \"Success\": \n",
    "            print(f\"\\t{check_json['constraint']} failed because: {check_json['constraint_message']}\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "conda_python3",
   "language": "python",
   "name": "conda_python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}