Spark SQL, DataFrames and Datasets Guide

Overview

Spark SQL is a Spark module for structured data processing. Unlike the basic Spark RDD API, the interfaces provided by Spark SQL provide Spark with more information about the structure of both the data and the computation being performed. Internally, Spark SQL uses this extra information to perform extra optimizations. There are several ways to interact with Spark SQL including SQL, the DataFrames API and the Datasets API. When computing a result the same execution engine is used, independent of which API/language you are using to express the computation. This unification means that developers can easily switch back and forth between the various APIs based on which provides the most natural way to express a given transformation.

All of the examples on this page use sample data included in the Spark distribution and can be run in the spark-shell, pyspark shell, or sparkR shell.

SQL

One use of Spark SQL is to execute SQL queries written using either a basic SQL syntax or HiveQL. Spark SQL can also be used to read data from an existing Hive installation. For more on how to configure this feature, please refer to the Hive Tables section. When running SQL from within another programming language the results will be returned as a DataFrame. You can also interact with the SQL interface using the command-line or over JDBC/ODBC.

DataFrames

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/Python, 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 RDDs.

The DataFrame API is available in Scala, Java, Python, and R.

Datasets

A Dataset is a new experimental interface added in Spark 1.6 that tries to provide the benefits of RDDs (strong typing, ability to use powerful lambda functions) with the benefits of Spark SQL’s optimized execution engine. A Dataset can be constructed from JVM objects and then manipulated using functional transformations (map, flatMap, filter, etc.).

The unified Dataset API can be used both in Scala and Java. Python does not yet have support for the Dataset API, but due to its dynamic nature many of the benefits are already available (i.e. you can access the field of a row by name naturally row.columnName). Full python support will be added in a future release.

Getting Started

Starting Point: SQLContext

The entry point into all functionality in Spark SQL is the SQLContext class, or one of its descendants. To create a basic SQLContext, all you need is a SparkContext.

val sc: SparkContext // An existing SparkContext.
val sqlContext = new org.apache.spark.sql.SQLContext(sc)

// this is used to implicitly convert an RDD to a DataFrame.
import sqlContext.implicits._

The entry point into all functionality in Spark SQL is the SQLContext class, or one of its descendants. To create a basic SQLContext, all you need is a SparkContext.

JavaSparkContext sc = ...; // An existing JavaSparkContext.
SQLContext sqlContext = new org.apache.spark.sql.SQLContext(sc);

The entry point into all relational functionality in Spark is the SQLContext class, or one of its decedents. To create a basic SQLContext, all you need is a SparkContext.

from pyspark.sql import SQLContext
sqlContext = SQLContext(sc)

The entry point into all relational functionality in Spark is the SQLContext class, or one of its decedents. To create a basic SQLContext, all you need is a SparkContext.

sqlContext <- sparkRSQL.init(sc)

In addition to the basic SQLContext, you can also create a HiveContext, which provides a superset of the functionality provided by the basic SQLContext. Additional features include the ability to write queries using the more complete HiveQL parser, access to Hive UDFs, and the ability to read data from Hive tables. To use a HiveContext, you do not need to have an existing Hive setup, and all of the data sources available to a SQLContext are still available. HiveContext is only packaged separately to avoid including all of Hive’s dependencies in the default Spark build. If these dependencies are not a problem for your application then using HiveContext is recommended for the 1.3 release of Spark. Future releases will focus on bringing SQLContext up to feature parity with a HiveContext.

Creating DataFrames

With a SQLContext, applications can create DataFrames from an existing RDD, from a Hive table, or from data sources.

As an example, the following creates a DataFrame based on the content of a JSON file:

val sc: SparkContext // An existing SparkContext.
val sqlContext = new org.apache.spark.sql.SQLContext(sc)

val df = sqlContext.read.json("examples/src/main/resources/people.json")

// Displays the content of the DataFrame to stdout
df.show()
JavaSparkContext sc = ...; // An existing JavaSparkContext.
SQLContext sqlContext = new org.apache.spark.sql.SQLContext(sc);

DataFrame df = sqlContext.read().json("examples/src/main/resources/people.json");

// Displays the content of the DataFrame to stdout
df.show();
from pyspark.sql import SQLContext
sqlContext = SQLContext(sc)

df = sqlContext.read.json("examples/src/main/resources/people.json")

# Displays the content of the DataFrame to stdout
df.show()
sqlContext <- SQLContext(sc)

df <- read.json(sqlContext, "examples/src/main/resources/people.json")

# Displays the content of the DataFrame to stdout
showDF(df)

DataFrame Operations

DataFrames provide a domain-specific language for structured data manipulation in Scala, Java, Python and R.

Here we include some basic examples of structured data processing using DataFrames:

val sc: SparkContext // An existing SparkContext.
val sqlContext = new org.apache.spark.sql.SQLContext(sc)

// Create the DataFrame
val df = sqlContext.read.json("examples/src/main/resources/people.json")

// Show the content of the DataFrame
df.show()
// age  name
// null Michael
// 30   Andy
// 19   Justin

// Print the schema in a tree format
df.printSchema()
// root
// |-- age: long (nullable = true)
// |-- name: string (nullable = true)

// Select only the "name" column
df.select("name").show()
// name
// Michael
// Andy
// Justin

// Select everybody, but increment the age by 1
df.select(df("name"), df("age") + 1).show()
// name    (age + 1)
// Michael null
// Andy    31
// Justin  20

// Select people older than 21
df.filter(df("age") > 21).show()
// age name
// 30  Andy

// Count people by age
df.groupBy("age").count().show()
// age  count
// null 1
// 19   1
// 30   1

For a complete list of the types of operations that can be performed on a DataFrame refer to the API Documentation.

In addition to simple column references and expressions, DataFrames also have a rich library of functions including string manipulation, date arithmetic, common math operations and more. The complete list is available in the DataFrame Function Reference.

JavaSparkContext sc // An existing SparkContext.
SQLContext sqlContext = new org.apache.spark.sql.SQLContext(sc)

// Create the DataFrame
DataFrame df = sqlContext.read().json("examples/src/main/resources/people.json");

// Show the content of the DataFrame
df.show();
// age  name
// null Michael
// 30   Andy
// 19   Justin

// Print the schema in a tree format
df.printSchema();
// root
// |-- age: long (nullable = true)
// |-- name: string (nullable = true)

// Select only the "name" column
df.select("name").show();
// name
// Michael
// Andy
// Justin

// Select everybody, but increment the age by 1
df.select(df.col("name"), df.col("age").plus(1)).show();
// name    (age + 1)
// Michael null
// Andy    31
// Justin  20

// Select people older than 21
df.filter(df.col("age").gt(21)).show();
// age name
// 30  Andy

// Count people by age
df.groupBy("age").count().show();
// age  count
// null 1
// 19   1
// 30   1

For a complete list of the types of operations that can be performed on a DataFrame refer to the API Documentation.

In addition to simple column references and expressions, DataFrames also have a rich library of functions including string manipulation, date arithmetic, common math operations and more. The complete list is available in the DataFrame Function Reference.

In Python it’s possible to access a DataFrame’s columns either by attribute (df.age) or by indexing (df['age']). While the former is convenient for interactive data exploration, users are highly encouraged to use the latter form, which is future proof and won’t break with column names that are also attributes on the DataFrame class.

from pyspark.sql import SQLContext
sqlContext = SQLContext(sc)

# Create the DataFrame
df = sqlContext.read.json("examples/src/main/resources/people.json")

# Show the content of the DataFrame
df.show()
## age  name
## null Michael
## 30   Andy
## 19   Justin

# Print the schema in a tree format
df.printSchema()
## root
## |-- age: long (nullable = true)
## |-- name: string (nullable = true)

# Select only the "name" column
df.select("name").show()
## name
## Michael
## Andy
## Justin

# Select everybody, but increment the age by 1
df.select(df['name'], df['age'] + 1).show()
## name    (age + 1)
## Michael null
## Andy    31
## Justin  20

# Select people older than 21
df.filter(df['age'] > 21).show()
## age name
## 30  Andy

# Count people by age
df.groupBy("age").count().show()
## age  count
## null 1
## 19   1
## 30   1

For a complete list of the types of operations that can be performed on a DataFrame refer to the API Documentation.

In addition to simple column references and expressions, DataFrames also have a rich library of functions including string manipulation, date arithmetic, common math operations and more. The complete list is available in the DataFrame Function Reference.

sqlContext <- sparkRSQL.init(sc)

# Create the DataFrame
df <- read.json(sqlContext, "examples/src/main/resources/people.json")

# Show the content of the DataFrame
showDF(df)
## age  name
## null Michael
## 30   Andy
## 19   Justin

# Print the schema in a tree format
printSchema(df)
## root
## |-- age: long (nullable = true)
## |-- name: string (nullable = true)

# Select only the "name" column
showDF(select(df, "name"))
## name
## Michael
## Andy
## Justin

# Select everybody, but increment the age by 1
showDF(select(df, df$name, df$age + 1))
## name    (age + 1)
## Michael null
## Andy    31
## Justin  20

# Select people older than 21
showDF(where(df, df$age > 21))
## age name
## 30  Andy

# Count people by age
showDF(count(groupBy(df, "age")))
## age  count
## null 1
## 19   1
## 30   1

For a complete list of the types of operations that can be performed on a DataFrame refer to the API Documentation.

In addition to simple column references and expressions, DataFrames also have a rich library of functions including string manipulation, date arithmetic, common math operations and more. The complete list is available in the DataFrame Function Reference.

Running SQL Queries Programmatically

The sql function on a SQLContext enables applications to run SQL queries programmatically and returns the result as a DataFrame.

val sqlContext = ... // An existing SQLContext
val df = sqlContext.sql("SELECT * FROM table")
SQLContext sqlContext = ... // An existing SQLContext
DataFrame df = sqlContext.sql("SELECT * FROM table")
from pyspark.sql import SQLContext
sqlContext = SQLContext(sc)
df = sqlContext.sql("SELECT * FROM table")
sqlContext <- sparkRSQL.init(sc)
df <- sql(sqlContext, "SELECT * FROM table")

Creating Datasets

Datasets are similar to RDDs, however, instead of using Java Serialization or Kryo they use a specialized Encoder to serialize the objects for processing or transmitting over the network. While both encoders and standard serialization are responsible for turning an object into bytes, encoders are code generated dynamically and use a format that allows Spark to perform many operations like filtering, sorting and hashing without deserializing the bytes back into an object.

// Encoders for most common types are automatically provided by importing sqlContext.implicits._
val ds = Seq(1, 2, 3).toDS()
ds.map(_ + 1).collect() // Returns: Array(2, 3, 4)

// Encoders are also created for case classes.
case class Person(name: String, age: Long)
val ds = Seq(Person("Andy", 32)).toDS()

// DataFrames can be converted to a Dataset by providing a class. Mapping will be done by name.
val path = "examples/src/main/resources/people.json"
val people = sqlContext.read.json(path).as[Person]
JavaSparkContext sc = ...; // An existing JavaSparkContext.
SQLContext sqlContext = new org.apache.spark.sql.SQLContext(sc);

Interoperating with RDDs

Spark SQL supports two different methods for converting existing RDDs into DataFrames. The first method uses reflection to infer the schema of an RDD that contains specific types of objects. This reflection based approach leads to more concise code and works well when you already know the schema while writing your Spark application.

The second method for creating DataFrames is through a programmatic interface that allows you to construct a schema and then apply it to an existing RDD. While this method is more verbose, it allows you to construct DataFrames when the columns and their types are not known until runtime.

Inferring the Schema Using Reflection

The Scala interface for Spark SQL supports automatically converting an RDD containing case classes to a DataFrame. The case class defines the schema of the table. The names of the arguments to the case class are read using reflection and become the names of the columns. Case classes can also be nested or contain complex types such as Sequences or Arrays. This RDD can be implicitly converted to a DataFrame and then be registered as a table. Tables can be used in subsequent SQL statements.

// sc is an existing SparkContext.
val sqlContext = new org.apache.spark.sql.SQLContext(sc)
// this is used to implicitly convert an RDD to a DataFrame.
import sqlContext.implicits._

// Define the schema using a case class.
// Note: Case classes in Scala 2.10 can support only up to 22 fields. To work around this limit,
// you can use custom classes that implement the Product interface.
case class Person(name: String, age: Int)

// Create an RDD of Person objects and register it as a table.
val people = sc.textFile("examples/src/main/resources/people.txt").map(_.split(",")).map(p => Person(p(0), p(1).trim.toInt)).toDF()
people.createOrReplaceTempView("people")

// SQL statements can be run by using the sql methods provided by sqlContext.
val teenagers = sqlContext.sql("SELECT name, age FROM people WHERE age >= 13 AND age <= 19")

// The results of SQL queries are DataFrames and support all the normal RDD operations.
// The columns of a row in the result can be accessed by field index:
teenagers.map(t => "Name: " + t(0)).collect().foreach(println)

// or by field name:
teenagers.map(t => "Name: " + t.getAs[String]("name")).collect().foreach(println)

// row.getValuesMap[T] retrieves multiple columns at once into a Map[String, T]
teenagers.map(_.getValuesMap[Any](List("name", "age"))).collect().foreach(println)
// Map("name" -> "Justin", "age" -> 19)

Spark SQL supports automatically converting an RDD of JavaBeans into a DataFrame. The BeanInfo, obtained using reflection, defines the schema of the table. Currently, Spark SQL does not support JavaBeans that contain nested or contain complex types such as Lists or Arrays. You can create a JavaBean by creating a class that implements Serializable and has getters and setters for all of its fields.

public static class Person implements Serializable {
  private String name;
  private int age;

  public String getName() {
    return name;
  }

  public void setName(String name) {
    this.name = name;
  }

  public int getAge() {
    return age;
  }

  public void setAge(int age) {
    this.age = age;
  }
}

A schema can be applied to an existing RDD by calling createDataFrame and providing the Class object for the JavaBean.

// sc is an existing JavaSparkContext.
SQLContext sqlContext = new org.apache.spark.sql.SQLContext(sc);

// Load a text file and convert each line to a JavaBean.
JavaRDD<Person> people = sc.textFile("examples/src/main/resources/people.txt").map(
  new Function<String, Person>() {
    public Person call(String line) throws Exception {
      String[] parts = line.split(",");

      Person person = new Person();
      person.setName(parts[0]);
      person.setAge(Integer.parseInt(parts[1].trim()));

      return person;
    }
  });

// Apply a schema to an RDD of JavaBeans and register it as a table.
DataFrame schemaPeople = sqlContext.createDataFrame(people, Person.class);
schemaPeople.createOrReplaceTempView("people");

// SQL can be run over RDDs that have been registered as tables.
DataFrame teenagers = sqlContext.sql("SELECT name FROM people WHERE age >= 13 AND age <= 19")

// The results of SQL queries are DataFrames and support all the normal RDD operations.
// The columns of a row in the result can be accessed by ordinal.
List<String> teenagerNames = teenagers.javaRDD().map(new Function<Row, String>() {
  public String call(Row row) {
    return "Name: " + row.getString(0);
  }
}).collect();

Spark SQL can convert an RDD of Row objects to a DataFrame, inferring the datatypes. Rows are constructed by passing a list of key/value pairs as kwargs to the Row class. The keys of this list define the column names of the table, and the types are inferred by looking at the first row. Since we currently only look at the first row, it is important that there is no missing data in the first row of the RDD. In future versions we plan to more completely infer the schema by looking at more data, similar to the inference that is performed on JSON files.

# sc is an existing SparkContext.
from pyspark.sql import SQLContext, Row
sqlContext = SQLContext(sc)

# Load a text file and convert each line to a Row.
lines = sc.textFile("examples/src/main/resources/people.txt")
parts = lines.map(lambda l: l.split(","))
people = parts.map(lambda p: Row(name=p[0], age=int(p[1])))

# Infer the schema, and register the DataFrame as a table.
schemaPeople = sqlContext.createDataFrame(people)
schemaPeople.createOrReplaceTempView("people")

# SQL can be run over DataFrames that have been registered as a table.
teenagers = sqlContext.sql("SELECT name FROM people WHERE age >= 13 AND age <= 19")

# The results of SQL queries are RDDs and support all the normal RDD operations.
teenNames = teenagers.map(lambda p: "Name: " + p.name)
for teenName in teenNames.collect():
  print(teenName)

Programmatically Specifying the Schema

When case classes cannot be defined ahead of time (for example, the structure of records is encoded in a string, or a text dataset will be parsed and fields will be projected differently for different users), a DataFrame can be created programmatically with three steps.

  1. Create an RDD of Rows from the original RDD;
  2. Create the schema represented by a StructType matching the structure of Rows in the RDD created in Step 1.
  3. Apply the schema to the RDD of Rows via createDataFrame method provided by SQLContext.

For example:

// sc is an existing SparkContext.
val sqlContext = new org.apache.spark.sql.SQLContext(sc)

// Create an RDD
val people = sc.textFile("examples/src/main/resources/people.txt")

// The schema is encoded in a string
val schemaString = "name age"

// Import Row.
import org.apache.spark.sql.Row;

// Import Spark SQL data types
import org.apache.spark.sql.types.{StructType,StructField,StringType};

// Generate the schema based on the string of schema
val schema =
  StructType(
    schemaString.split(" ").map(fieldName => StructField(fieldName, StringType, true)))

// Convert records of the RDD (people) to Rows.
val rowRDD = people.map(_.split(",")).map(p => Row(p(0), p(1).trim))

// Apply the schema to the RDD.
val peopleDataFrame = sqlContext.createDataFrame(rowRDD, schema)

// Creates a temporary view using the DataFrame.
peopleDataFrame.createOrReplaceTempView("people")

// SQL statements can be run by using the sql methods provided by sqlContext.
val results = sqlContext.sql("SELECT name FROM people")

// The results of SQL queries are DataFrames and support all the normal RDD operations.
// The columns of a row in the result can be accessed by field index or by field name.
results.map(t => "Name: " + t(0)).collect().foreach(println)

When JavaBean classes cannot be defined ahead of time (for example, the structure of records is encoded in a string, or a text dataset will be parsed and fields will be projected differently for different users), a DataFrame can be created programmatically with three steps.

  1. Create an RDD of Rows from the original RDD;
  2. Create the schema represented by a StructType matching the structure of Rows in the RDD created in Step 1.
  3. Apply the schema to the RDD of Rows via createDataFrame method provided by SQLContext.

For example:

import org.apache.spark.api.java.function.Function;
// Import factory methods provided by DataTypes.
import org.apache.spark.sql.types.DataTypes;
// Import StructType and StructField
import org.apache.spark.sql.types.StructType;
import org.apache.spark.sql.types.StructField;
// Import Row.
import org.apache.spark.sql.Row;
// Import RowFactory.
import org.apache.spark.sql.RowFactory;

// sc is an existing JavaSparkContext.
SQLContext sqlContext = new org.apache.spark.sql.SQLContext(sc);

// Load a text file and convert each line to a JavaBean.
JavaRDD<String> people = sc.textFile("examples/src/main/resources/people.txt");

// The schema is encoded in a string
String schemaString = "name age";

// Generate the schema based on the string of schema
List<StructField> fields = new ArrayList<>();
for (String fieldName: schemaString.split(" ")) {
  fields.add(DataTypes.createStructField(fieldName, DataTypes.StringType, true));
}
StructType schema = DataTypes.createStructType(fields);

// Convert records of the RDD (people) to Rows.
JavaRDD<Row> rowRDD = people.map(
  new Function<String, Row>() {
    public Row call(String record) throws Exception {
      String[] fields = record.split(",");
      return RowFactory.create(fields[0], fields[1].trim());
    }
  });

// Apply the schema to the RDD.
DataFrame peopleDataFrame = sqlContext.createDataFrame(rowRDD, schema);

// Creates a temporary view using the DataFrame.
peopleDataFrame.createOrReplaceTempView("people");

// SQL can be run over a temporary view created using DataFrames.
DataFrame results = sqlContext.sql("SELECT name FROM people");

// The results of SQL queries are DataFrames and support all the normal RDD operations.
// The columns of a row in the result can be accessed by ordinal.
List<String> names = results.javaRDD().map(new Function<Row, String>() {
  public String call(Row row) {
    return "Name: " + row.getString(0);
  }
}).collect();

When a dictionary of kwargs cannot be defined ahead of time (for example, the structure of records is encoded in a string, or a text dataset will be parsed and fields will be projected differently for different users), a DataFrame can be created programmatically with three steps.

  1. Create an RDD of tuples or lists from the original RDD;
  2. Create the schema represented by a StructType matching the structure of tuples or lists in the RDD created in the step 1.
  3. Apply the schema to the RDD via createDataFrame method provided by SQLContext.

For example:

# Import SQLContext and data types
from pyspark.sql import SQLContext
from pyspark.sql.types import *

# sc is an existing SparkContext.
sqlContext = SQLContext(sc)

# Load a text file and convert each line to a tuple.
lines = sc.textFile("examples/src/main/resources/people.txt")
parts = lines.map(lambda l: l.split(","))
people = parts.map(lambda p: (p[0], p[1].strip()))

# The schema is encoded in a string.
schemaString = "name age"

fields = [StructField(field_name, StringType(), True) for field_name in schemaString.split()]
schema = StructType(fields)

# Apply the schema to the RDD.
schemaPeople = sqlContext.createDataFrame(people, schema)

# Creates a temporary view using the DataFrame
schemaPeople.createOrReplaceTempView("people")

# SQL can be run over DataFrames that have been registered as a table.
results = sqlContext.sql("SELECT name FROM people")

# The results of SQL queries are RDDs and support all the normal RDD operations.
names = results.map(lambda p: "Name: " + p.name)
for name in names.collect():
  print(name)

Data Sources

Spark SQL supports operating on a variety of data sources through the DataFrame interface. A DataFrame can be operated on as normal RDDs and can also be used to create a temporary view. Registering a DataFrame as a table allows you to run SQL queries over its data. This section describes the general methods for loading and saving data using the Spark Data Sources and then goes into specific options that are available for the built-in data sources.

Generic Load/Save Functions

In the simplest form, the default data source (parquet unless otherwise configured by spark.sql.sources.default) will be used for all operations.

val df = sqlContext.read.load("examples/src/main/resources/users.parquet")
df.select("name", "favorite_color").write.save("namesAndFavColors.parquet")
DataFrame df = sqlContext.read().load("examples/src/main/resources/users.parquet");
df.select("name", "favorite_color").write().save("namesAndFavColors.parquet");
df = sqlContext.read.load("examples/src/main/resources/users.parquet")
df.select("name", "favorite_color").write.save("namesAndFavColors.parquet")
df <- read.df(sqlContext, "examples/src/main/resources/users.parquet")
write.df(select(df, "name", "favorite_color"), "namesAndFavColors.parquet")

Manually Specifying Options

You can also manually specify the data source that will be used along with any extra options that you would like to pass to the data source. Data sources are specified by their fully qualified name (i.e., org.apache.spark.sql.parquet), but for built-in sources you can also use their short names (json, parquet, jdbc). DataFrames of any type can be converted into other types using this syntax.

val df = sqlContext.read.format("json").load("examples/src/main/resources/people.json")
df.select("name", "age").write.format("parquet").save("namesAndAges.parquet")
DataFrame df = sqlContext.read().format("json").load("examples/src/main/resources/people.json");
df.select("name", "age").write().format("parquet").save("namesAndAges.parquet");
df = sqlContext.read.load("examples/src/main/resources/people.json", format="json")
df.select("name", "age").write.save("namesAndAges.parquet", format="parquet")
df <- read.df(sqlContext, "examples/src/main/resources/people.json", "json")
write.df(select(df, "name", "age"), "namesAndAges.parquet", "parquet")

Run SQL on files directly

Instead of using read API to load a file into DataFrame and query it, you can also query that file directly with SQL.

val df = sqlContext.sql("SELECT * FROM parquet.`examples/src/main/resources/users.parquet`")
DataFrame df = sqlContext.sql("SELECT * FROM parquet.`examples/src/main/resources/users.parquet`");
df = sqlContext.sql("SELECT * FROM parquet.`examples/src/main/resources/users.parquet`")
df <- sql(sqlContext, "SELECT * FROM parquet.`examples/src/main/resources/users.parquet`")

Save Modes

Save operations can optionally take a SaveMode, that specifies how to handle existing data if present. It is important to realize that these save modes do not utilize any locking and are not atomic. Additionally, when performing a Overwrite, the data will be deleted before writing out the new data.

Scala/JavaAny LanguageMeaning
SaveMode.ErrorIfExists (default) "error" (default) When saving a DataFrame to a data source, if data already exists, an exception is expected to be thrown.
SaveMode.Append "append" When saving a DataFrame to a data source, if data/table already exists, contents of the DataFrame are expected to be appended to existing data.
SaveMode.Overwrite "overwrite" Overwrite mode means that when saving a DataFrame to a data source, if data/table already exists, existing data is expected to be overwritten by the contents of the DataFrame.
SaveMode.Ignore "ignore" Ignore mode means that when saving a DataFrame to a data source, if data already exists, the save operation is expected to not save the contents of the DataFrame and to not change the existing data. This is similar to a CREATE TABLE IF NOT EXISTS in SQL.

Saving to Persistent Tables

When working with a HiveContext, DataFrames can also be saved as persistent tables using the saveAsTable command. Unlike the registerTempTable command, saveAsTable will materialize the contents of the dataframe and create a pointer to the data in the HiveMetastore. Persistent tables will still exist even after your Spark program has restarted, as long as you maintain your connection to the same metastore. A DataFrame for a persistent table can be created by calling the table method on a SQLContext with the name of the table.

By default saveAsTable will create a “managed table”, meaning that the location of the data will be controlled by the metastore. Managed tables will also have their data deleted automatically when a table is dropped.

Parquet Files

Parquet is a columnar format that is supported by many other data processing systems. Spark SQL provides support for both reading and writing Parquet files that automatically preserves the schema of the original data. When writing Parquet files, all columns are automatically converted to be nullable for compatibility reasons.

Loading Data Programmatically

Using the data from the above example:

// sqlContext from the previous example is used in this example.
// This is used to implicitly convert an RDD to a DataFrame.
import sqlContext.implicits._

val people: RDD[Person] = ... // An RDD of case class objects, from the previous example.

// The RDD is implicitly converted to a DataFrame by implicits, allowing it to be stored using Parquet.
people.write.parquet("people.parquet")

// Read in the parquet file created above. Parquet files are self-describing so the schema is preserved.
// The result of loading a Parquet file is also a DataFrame.
val parquetFile = sqlContext.read.parquet("people.parquet")

// Parquet files can also be used to create a temporary view and then used in SQL statements.
parquetFile.createOrReplaceTempView("parquetFile")
val teenagers = sqlContext.sql("SELECT name FROM parquetFile WHERE age >= 13 AND age <= 19")
teenagers.map(t => "Name: " + t(0)).collect().foreach(println)
// sqlContext from the previous example is used in this example.

DataFrame schemaPeople = ... // The DataFrame from the previous example.

// DataFrames can be saved as Parquet files, maintaining the schema information.
schemaPeople.write().parquet("people.parquet");

// Read in the Parquet file created above. Parquet files are self-describing so the schema is preserved.
// The result of loading a parquet file is also a DataFrame.
DataFrame parquetFile = sqlContext.read().parquet("people.parquet");

// Parquet files can also be used to create a temporary view and then used in SQL statements.
parquetFile.createOrReplaceTempView("parquetFile");
DataFrame teenagers = sqlContext.sql("SELECT name FROM parquetFile WHERE age >= 13 AND age <= 19");
List<String> teenagerNames = teenagers.javaRDD().map(new Function<Row, String>() {
  public String call(Row row) {
    return "Name: " + row.getString(0);
  }
}).collect();
# sqlContext from the previous example is used in this example.

schemaPeople # The DataFrame from the previous example.

# DataFrames can be saved as Parquet files, maintaining the schema information.
schemaPeople.write.parquet("people.parquet")

# Read in the Parquet file created above. Parquet files are self-describing so the schema is preserved.
# The result of loading a parquet file is also a DataFrame.
parquetFile = sqlContext.read.parquet("people.parquet")

# Parquet files can also be used to create a temporary view and then used in SQL statements.
parquetFile.createOrReplaceTempView("parquetFile");
teenagers = sqlContext.sql("SELECT name FROM parquetFile WHERE age >= 13 AND age <= 19")
teenNames = teenagers.map(lambda p: "Name: " + p.name)
for teenName in teenNames.collect():
  print(teenName)
# sqlContext from the previous example is used in this example.

schemaPeople # The DataFrame from the previous example.

# DataFrames can be saved as Parquet files, maintaining the schema information.
write.parquet(schemaPeople, "people.parquet")

# Read in the Parquet file created above. Parquet files are self-describing so the schema is preserved.
# The result of loading a parquet file is also a DataFrame.
parquetFile <- read.parquet(sqlContext, "people.parquet")

# Parquet files can also be used to create a temporary view and then used in SQL statements.
registerTempTable(parquetFile, "parquetFile")
teenagers <- sql(sqlContext, "SELECT name FROM parquetFile WHERE age >= 13 AND age <= 19")
schema <- structType(structField("name", "string"))
teenNames <- dapply(df, function(p) { cbind(paste("Name:", p$name)) }, schema)
for (teenName in collect(teenNames)$name) {
  cat(teenName, "\n")
}
CREATE TEMPORARY TABLE parquetTable
USING org.apache.spark.sql.parquet
OPTIONS (
  path "examples/src/main/resources/people.parquet"
)

SELECT * FROM parquetTable

Partition Discovery

Table partitioning is a common optimization approach used in systems like Hive. In a partitioned table, data are usually stored in different directories, with partitioning column values encoded in the path of each partition directory. The Parquet data source is now able to discover and infer partitioning information automatically. For example, we can store all our previously used population data into a partitioned table using the following directory structure, with two extra columns, gender and country as partitioning columns:

path
└── to
    └── table
        ├── gender=male
        │   ├── ...
        │   │
        │   ├── country=US
        │   │   └── data.parquet
        │   ├── country=CN
        │   │   └── data.parquet
        │   └── ...
        └── gender=female
            ├── ...
            │
            ├── country=US
            │   └── data.parquet
            ├── country=CN
            │   └── data.parquet
            └── ...

By passing path/to/table to either SQLContext.read.parquet or SQLContext.read.load, Spark SQL will automatically extract the partitioning information from the paths. Now the schema of the returned DataFrame becomes:

root
|-- name: string (nullable = true)
|-- age: long (nullable = true)
|-- gender: string (nullable = true)
|-- country: string (nullable = true)

Notice that the data types of the partitioning columns are automatically inferred. Currently, numeric data types and string type are supported. Sometimes users may not want to automatically infer the data types of the partitioning columns. For these use cases, the automatic type inference can be configured by spark.sql.sources.partitionColumnTypeInference.enabled, which is default to true. When type inference is disabled, string type will be used for the partitioning columns.

Starting from Spark 1.6.0, partition discovery only finds partitions under the given paths by default. For the above example, if users pass path/to/table/gender=male to either SQLContext.read.parquet or SQLContext.read.load, gender will not be considered as a partitioning column. If users need to specify the base path that partition discovery should start with, they can set basePath in the data source options. For example, when path/to/table/gender=male is the path of the data and users set basePath to path/to/table/, gender will be a partitioning column.

Schema Merging

Like ProtocolBuffer, Avro, and Thrift, Parquet also supports schema evolution. Users can start with a simple schema, and gradually add more columns to the schema as needed. In this way, users may end up with multiple Parquet files with different but mutually compatible schemas. The Parquet data source is now able to automatically detect this case and merge schemas of all these files.

Since schema merging is a relatively expensive operation, and is not a necessity in most cases, we turned it off by default starting from 1.5.0. You may enable it by

  1. setting data source option mergeSchema to true when reading Parquet files (as shown in the examples below), or
  2. setting the global SQL option spark.sql.parquet.mergeSchema to true.
// sqlContext from the previous example is used in this example.
// This is used to implicitly convert an RDD to a DataFrame.
import sqlContext.implicits._

// Create a simple DataFrame, stored into a partition directory
val df1 = sc.makeRDD(1 to 5).map(i => (i, i * 2)).toDF("single", "double")
df1.write.parquet("data/test_table/key=1")

// Create another DataFrame in a new partition directory,
// adding a new column and dropping an existing column
val df2 = sc.makeRDD(6 to 10).map(i => (i, i * 3)).toDF("single", "triple")
df2.write.parquet("data/test_table/key=2")

// Read the partitioned table
val df3 = sqlContext.read.option("mergeSchema", "true").parquet("data/test_table")
df3.printSchema()

// The final schema consists of all 3 columns in the Parquet files together
// with the partitioning column appeared in the partition directory paths.
// root
// |-- single: int (nullable = true)
// |-- double: int (nullable = true)
// |-- triple: int (nullable = true)
// |-- key : int (nullable = true)
# sqlContext from the previous example is used in this example.

# Create a simple DataFrame, stored into a partition directory
df1 = sqlContext.createDataFrame(sc.parallelize(range(1, 6))\
                                   .map(lambda i: Row(single=i, double=i * 2)))
df1.write.parquet("data/test_table/key=1")

# Create another DataFrame in a new partition directory,
# adding a new column and dropping an existing column
df2 = sqlContext.createDataFrame(sc.parallelize(range(6, 11))
                                   .map(lambda i: Row(single=i, triple=i * 3)))
df2.write.parquet("data/test_table/key=2")

# Read the partitioned table
df3 = sqlContext.read.option("mergeSchema", "true").parquet("data/test_table")
df3.printSchema()

# The final schema consists of all 3 columns in the Parquet files together
# with the partitioning column appeared in the partition directory paths.
# root
# |-- single: int (nullable = true)
# |-- double: int (nullable = true)
# |-- triple: int (nullable = true)
# |-- key : int (nullable = true)
# sqlContext from the previous example is used in this example.

# Create a simple DataFrame, stored into a partition directory
write.df(df1, "data/test_table/key=1", "parquet", "overwrite")

# Create another DataFrame in a new partition directory,
# adding a new column and dropping an existing column
write.df(df2, "data/test_table/key=2", "parquet", "overwrite")

# Read the partitioned table
df3 <- read.df(sqlContext, "data/test_table", "parquet", mergeSchema="true")
printSchema(df3)

# The final schema consists of all 3 columns in the Parquet files together
# with the partitioning column appeared in the partition directory paths.
# root
# |-- single: int (nullable = true)
# |-- double: int (nullable = true)
# |-- triple: int (nullable = true)
# |-- key : int (nullable = true)

Hive metastore Parquet table conversion

When reading from and writing to Hive metastore Parquet tables, Spark SQL will try to use its own Parquet support instead of Hive SerDe for better performance. This behavior is controlled by the spark.sql.hive.convertMetastoreParquet configuration, and is turned on by default.

Hive/Parquet Schema Reconciliation

There are two key differences between Hive and Parquet from the perspective of table schema processing.

  1. Hive is case insensitive, while Parquet is not
  2. Hive considers all columns nullable, while nullability in Parquet is significant

Due to this reason, we must reconcile Hive metastore schema with Parquet schema when converting a Hive metastore Parquet table to a Spark SQL Parquet table. The reconciliation rules are:

  1. Fields that have the same name in both schema must have the same data type regardless of nullability. The reconciled field should have the data type of the Parquet side, so that nullability is respected.

  2. The reconciled schema contains exactly those fields defined in Hive metastore schema.

    • Any fields that only appear in the Parquet schema are dropped in the reconciled schema.
    • Any fields that only appear in the Hive metastore schema are added as nullable field in the reconciled schema.

Metadata Refreshing

Spark SQL caches Parquet metadata for better performance. When Hive metastore Parquet table conversion is enabled, metadata of those converted tables are also cached. If these tables are updated by Hive or other external tools, you need to refresh them manually to ensure consistent metadata.

// sqlContext is an existing HiveContext
sqlContext.refreshTable("my_table")
// sqlContext is an existing HiveContext
sqlContext.refreshTable("my_table")
# sqlContext is an existing HiveContext
sqlContext.refreshTable("my_table")
REFRESH TABLE my_table;

Configuration

Configuration of Parquet can be done using the setConf method on SQLContext or by running SET key=value commands using SQL.

Property NameDefaultMeaning
spark.sql.parquet.binaryAsString false Some other Parquet-producing systems, in particular Impala, Hive, and older versions of Spark SQL, do not differentiate between binary data and strings when writing out the Parquet schema. This flag tells Spark SQL to interpret binary data as a string to provide compatibility with these systems.
spark.sql.parquet.int96AsTimestamp true Some Parquet-producing systems, in particular Impala and Hive, store Timestamp into INT96. This flag tells Spark SQL to interpret INT96 data as a timestamp to provide compatibility with these systems.
spark.sql.parquet.cacheMetadata true Turns on caching of Parquet schema metadata. Can speed up querying of static data.
spark.sql.parquet.compression.codec gzip Sets the compression codec use when writing Parquet files. Acceptable values include: uncompressed, snappy, gzip, lzo.
spark.sql.parquet.filterPushdown true Enables Parquet filter push-down optimization when set to true.
spark.sql.hive.convertMetastoreParquet true When set to false, Spark SQL will use the Hive SerDe for parquet tables instead of the built in support.
spark.sql.parquet.mergeSchema false

When true, the Parquet data source merges schemas collected from all data files, otherwise the schema is picked from the summary file or a random data file if no summary file is available.

JSON Datasets

Spark SQL can automatically infer the schema of a JSON dataset and load it as a DataFrame. This conversion can be done using SQLContext.read.json() on either an RDD of String, or a JSON file.

Note that the file that is offered as a json file is not a typical JSON file. Each line must contain a separate, self-contained valid JSON object. As a consequence, a regular multi-line JSON file will most often fail.

// sc is an existing SparkContext.
val sqlContext = new org.apache.spark.sql.SQLContext(sc)

// A JSON dataset is pointed to by path.
// The path can be either a single text file or a directory storing text files.
val path = "examples/src/main/resources/people.json"
val people = sqlContext.read.json(path)

// The inferred schema can be visualized using the printSchema() method.
people.printSchema()
// root
//  |-- age: long (nullable = true)
//  |-- name: string (nullable = true)

// Creates a temporary view using the DataFrame
people.createOrReplaceTempView("people")

// SQL statements can be run by using the sql methods provided by sqlContext.
val teenagers = sqlContext.sql("SELECT name FROM people WHERE age >= 13 AND age <= 19")

// Alternatively, a DataFrame can be created for a JSON dataset represented by
// an RDD[String] storing one JSON object per string.
val anotherPeopleRDD = sc.parallelize(
  """{"name":"Yin","address":{"city":"Columbus","state":"Ohio"}}""" :: Nil)
val anotherPeople = sqlContext.read.json(anotherPeopleRDD)

Spark SQL can automatically infer the schema of a JSON dataset and load it as a DataFrame. This conversion can be done using SQLContext.read().json() on either an RDD of String, or a JSON file.

Note that the file that is offered as a json file is not a typical JSON file. Each line must contain a separate, self-contained valid JSON object. As a consequence, a regular multi-line JSON file will most often fail.

// sc is an existing JavaSparkContext.
SQLContext sqlContext = new org.apache.spark.sql.SQLContext(sc);

// A JSON dataset is pointed to by path.
// The path can be either a single text file or a directory storing text files.
DataFrame people = sqlContext.read().json("examples/src/main/resources/people.json");

// The inferred schema can be visualized using the printSchema() method.
people.printSchema();
// root
//  |-- age: long (nullable = true)
//  |-- name: string (nullable = true)

// Creates a temporary view using the DataFrame
people.createOrReplaceTempView("people");

// SQL statements can be run by using the sql methods provided by sqlContext.
DataFrame teenagers = sqlContext.sql("SELECT name FROM people WHERE age >= 13 AND age <= 19");

// Alternatively, a DataFrame can be created for a JSON dataset represented by
// an RDD[String] storing one JSON object per string.
List<String> jsonData = Arrays.asList(
  "{\"name\":\"Yin\",\"address\":{\"city\":\"Columbus\",\"state\":\"Ohio\"}}");
JavaRDD<String> anotherPeopleRDD = sc.parallelize(jsonData);
DataFrame anotherPeople = sqlContext.read().json(anotherPeopleRDD);

Spark SQL can automatically infer the schema of a JSON dataset and load it as a DataFrame. This conversion can be done using SQLContext.read.json on a JSON file.

Note that the file that is offered as a json file is not a typical JSON file. Each line must contain a separate, self-contained valid JSON object. As a consequence, a regular multi-line JSON file will most often fail.

# sc is an existing SparkContext.
from pyspark.sql import SQLContext
sqlContext = SQLContext(sc)

# A JSON dataset is pointed to by path.
# The path can be either a single text file or a directory storing text files.
people = sqlContext.read.json("examples/src/main/resources/people.json")

# The inferred schema can be visualized using the printSchema() method.
people.printSchema()
# root
#  |-- age: long (nullable = true)
#  |-- name: string (nullable = true)

# Creates a temporary view using the DataFrame.
people.createOrReplaceTempView("people")

# SQL statements can be run by using the sql methods provided by `sqlContext`.
teenagers = sqlContext.sql("SELECT name FROM people WHERE age >= 13 AND age <= 19")

# Alternatively, a DataFrame can be created for a JSON dataset represented by
# an RDD[String] storing one JSON object per string.
anotherPeopleRDD = sc.parallelize([
  '{"name":"Yin","address":{"city":"Columbus","state":"Ohio"}}'])
anotherPeople = sqlContext.jsonRDD(anotherPeopleRDD)

Spark SQL can automatically infer the schema of a JSON dataset and load it as a DataFrame. using the jsonFile function, which loads data from a directory of JSON files where each line of the files is a JSON object.

Note that the file that is offered as a json file is not a typical JSON file. Each line must contain a separate, self-contained valid JSON object. As a consequence, a regular multi-line JSON file will most often fail.

# sc is an existing SparkContext.
sqlContext <- sparkRSQL.init(sc)

# A JSON dataset is pointed to by path.
# The path can be either a single text file or a directory storing text files.
path <- "examples/src/main/resources/people.json"
# Create a DataFrame from the file(s) pointed to by path
people <- read.json(sqlContext, path)

# The inferred schema can be visualized using the printSchema() method.
printSchema(people)
# root
#  |-- age: long (nullable = true)
#  |-- name: string (nullable = true)

# Register this DataFrame as a table.
registerTempTable(people, "people")

# SQL statements can be run by using the sql methods provided by `sqlContext`.
teenagers <- sql(sqlContext, "SELECT name FROM people WHERE age >= 13 AND age <= 19")
CREATE TEMPORARY TABLE jsonTable
USING org.apache.spark.sql.json
OPTIONS (
  path "examples/src/main/resources/people.json"
)

SELECT * FROM jsonTable

Hive Tables

Spark SQL also supports reading and writing data stored in Apache Hive. However, since Hive has a large number of dependencies, it is not included in the default Spark assembly. Hive support is enabled by adding the -Phive and -Phive-thriftserver flags to Spark’s build. This command builds a new assembly directory that includes Hive. Note that this Hive assembly directory must also be present on all of the worker nodes, as they will need access to the Hive serialization and deserialization libraries (SerDes) in order to access data stored in Hive.

Configuration of Hive is done by placing your hive-site.xml, core-site.xml (for security configuration), hdfs-site.xml (for HDFS configuration) file in conf/.

When working with Hive one must construct a HiveContext, which inherits from SQLContext, and adds support for finding tables in the MetaStore and writing queries using HiveQL. Users who do not have an existing Hive deployment can still create a HiveContext. When not configured by the hive-site.xml, the context automatically creates metastore_db in the current directory and creates warehouse directory indicated by HiveConf, which defaults to /user/hive/warehouse. Note that you may need to grant write privilege on /user/hive/warehouse to the user who starts the spark application.

// sc is an existing SparkContext.
val sqlContext = new org.apache.spark.sql.hive.HiveContext(sc)

sqlContext.sql("CREATE TABLE IF NOT EXISTS src (key INT, value STRING)")
sqlContext.sql("LOAD DATA LOCAL INPATH 'examples/src/main/resources/kv1.txt' INTO TABLE src")

// Queries are expressed in HiveQL
sqlContext.sql("FROM src SELECT key, value").collect().foreach(println)

When working with Hive one must construct a HiveContext, which inherits from SQLContext, and adds support for finding tables in the MetaStore and writing queries using HiveQL. In addition to the sql method a HiveContext also provides an hql method, which allows queries to be expressed in HiveQL.

// sc is an existing JavaSparkContext.
HiveContext sqlContext = new org.apache.spark.sql.hive.HiveContext(sc.sc);

sqlContext.sql("CREATE TABLE IF NOT EXISTS src (key INT, value STRING)");
sqlContext.sql("LOAD DATA LOCAL INPATH 'examples/src/main/resources/kv1.txt' INTO TABLE src");

// Queries are expressed in HiveQL.
Row[] results = sqlContext.sql("FROM src SELECT key, value").collect();

When working with Hive one must construct a HiveContext, which inherits from SQLContext, and adds support for finding tables in the MetaStore and writing queries using HiveQL.

# sc is an existing SparkContext.
from pyspark.sql import HiveContext
sqlContext = HiveContext(sc)

sqlContext.sql("CREATE TABLE IF NOT EXISTS src (key INT, value STRING)")
sqlContext.sql("LOAD DATA LOCAL INPATH 'examples/src/main/resources/kv1.txt' INTO TABLE src")

# Queries can be expressed in HiveQL.
results = sqlContext.sql("FROM src SELECT key, value").collect()

When working with Hive one must construct a HiveContext, which inherits from SQLContext, and adds support for finding tables in the MetaStore and writing queries using HiveQL.

# sc is an existing SparkContext.
sqlContext <- sparkRHive.init(sc)

sql(sqlContext, "CREATE TABLE IF NOT EXISTS src (key INT, value STRING)")
sql(sqlContext, "LOAD DATA LOCAL INPATH 'examples/src/main/resources/kv1.txt' INTO TABLE src")

# Queries can be expressed in HiveQL.
results <- collect(sql(sqlContext, "FROM src SELECT key, value"))

Interacting with Different Versions of Hive Metastore

One of the most important pieces of Spark SQL’s Hive support is interaction with Hive metastore, which enables Spark SQL to access metadata of Hive tables. Starting from Spark 1.4.0, a single binary build of Spark SQL can be used to query different versions of Hive metastores, using the configuration described below. Note that independent of the version of Hive that is being used to talk to the metastore, internally Spark SQL will compile against Hive 1.2.1 and use those classes for internal execution (serdes, UDFs, UDAFs, etc).

The following options can be used to configure the version of Hive that is used to retrieve metadata:

Property NameDefaultMeaning
spark.sql.hive.metastore.version 1.2.1 Version of the Hive metastore. Available options are 0.12.0 through 1.2.1.
spark.sql.hive.metastore.jars builtin Location of the jars that should be used to instantiate the HiveMetastoreClient. This property can be one of three options:
  1. builtin
  2. Use Hive 1.2.1, which is bundled with the Spark assembly when -Phive is enabled. When this option is chosen, spark.sql.hive.metastore.version must be either 1.2.1 or not defined.
  3. maven
  4. Use Hive jars of specified version downloaded from Maven repositories. This configuration is not generally recommended for production deployments.
  5. A classpath in the standard format for the JVM. This classpath must include all of Hive and its dependencies, including the correct version of Hadoop. These jars only need to be present on the driver, but if you are running in yarn cluster mode then you must ensure they are packaged with you application.
spark.sql.hive.metastore.sharedPrefixes com.mysql.jdbc,
org.postgresql,
com.microsoft.sqlserver,
oracle.jdbc

A comma separated list of class prefixes that should be loaded using the classloader that is shared between Spark SQL and a specific version of Hive. An example of classes that should be shared is JDBC drivers that are needed to talk to the metastore. Other classes that need to be shared are those that interact with classes that are already shared. For example, custom appenders that are used by log4j.

spark.sql.hive.metastore.barrierPrefixes (empty)

A comma separated list of class prefixes that should explicitly be reloaded for each version of Hive that Spark SQL is communicating with. For example, Hive UDFs that are declared in a prefix that typically would be shared (i.e. org.apache.spark.*).

JDBC To Other Databases

Spark SQL also includes a data source that can read data from other databases using JDBC. This functionality should be preferred over using JdbcRDD. This is because the results are returned as a DataFrame and they can easily be processed in Spark SQL or joined with other data sources. The JDBC data source is also easier to use from Java or Python as it does not require the user to provide a ClassTag. (Note that this is different than the Spark SQL JDBC server, which allows other applications to run queries using Spark SQL).

To get started you will need to include the JDBC driver for you particular database on the spark classpath. For example, to connect to postgres from the Spark Shell you would run the following command:

bin/spark-shell --driver-class-path postgresql-9.4.1207.jar --jars postgresql-9.4.1207.jar

Tables from the remote database can be loaded as a DataFrame or Spark SQL Temporary table using the Data Sources API. The following options are supported:

Property NameMeaning
url The JDBC URL to connect to.
dbtable The JDBC table that should be read. Note that anything that is valid in a FROM clause of a SQL query can be used. For example, instead of a full table you could also use a subquery in parentheses.
driver The class name of the JDBC driver to use to connect to this URL.
partitionColumn, lowerBound, upperBound, numPartitions These options must all be specified if any of them is specified. They describe how to partition the table when reading in parallel from multiple workers. partitionColumn must be a numeric column from the table in question. Notice that lowerBound and upperBound are just used to decide the partition stride, not for filtering the rows in table. So all rows in the table will be partitioned and returned.
fetchSize The JDBC fetch size, which determines how many rows to fetch per round trip. This can help performance on JDBC drivers which default to low fetch size (eg. Oracle with 10 rows).
val jdbcDF = sqlContext.read.format("jdbc").options(
  Map("url" -> "jdbc:postgresql:dbserver",
  "dbtable" -> "schema.tablename")).load()
Map<String, String> options = new HashMap<>();
options.put("url", "jdbc:postgresql:dbserver");
options.put("dbtable", "schema.tablename");

DataFrame jdbcDF = sqlContext.read().format("jdbc"). options(options).load();
df = sqlContext.read.format('jdbc').options(url='jdbc:postgresql:dbserver', dbtable='schema.tablename').load()
df <- loadDF(sqlContext, source="jdbc", url="jdbc:postgresql:dbserver", dbtable="schema.tablename")
CREATE TEMPORARY TABLE jdbcTable
USING org.apache.spark.sql.jdbc
OPTIONS (
  url "jdbc:postgresql:dbserver",
  dbtable "schema.tablename"
)

Troubleshooting

Performance Tuning

For some workloads it is possible to improve performance by either caching data in memory, or by turning on some experimental options.

Caching Data In Memory

Spark SQL can cache tables using an in-memory columnar format by calling sqlContext.cacheTable("tableName") or dataFrame.cache(). Then Spark SQL will scan only required columns and will automatically tune compression to minimize memory usage and GC pressure. You can call sqlContext.uncacheTable("tableName") to remove the table from memory.

Configuration of in-memory caching can be done using the setConf method on SQLContext or by running SET key=value commands using SQL.

Property NameDefaultMeaning
spark.sql.inMemoryColumnarStorage.compressed true When set to true Spark SQL will automatically select a compression codec for each column based on statistics of the data.
spark.sql.inMemoryColumnarStorage.batchSize 10000 Controls the size of batches for columnar caching. Larger batch sizes can improve memory utilization and compression, but risk OOMs when caching data.

Other Configuration Options

The following options can also be used to tune the performance of query execution. It is possible that these options will be deprecated in future release as more optimizations are performed automatically.

Property NameDefaultMeaning
spark.sql.autoBroadcastJoinThreshold 10485760 (10 MB) Configures the maximum size in bytes for a table that will be broadcast to all worker nodes when performing a join. By setting this value to -1 broadcasting can be disabled. Note that currently statistics are only supported for Hive Metastore tables where the command ANALYZE TABLE <tableName> COMPUTE STATISTICS noscan has been run.
spark.sql.tungsten.enabled true When true, use the optimized Tungsten physical execution backend which explicitly manages memory and dynamically generates bytecode for expression evaluation.
spark.sql.shuffle.partitions 200 Configures the number of partitions to use when shuffling data for joins or aggregations.

Distributed SQL Engine

Spark SQL can also act as a distributed query engine using its JDBC/ODBC or command-line interface. In this mode, end-users or applications can interact with Spark SQL directly to run SQL queries, without the need to write any code.

Running the Thrift JDBC/ODBC server

The Thrift JDBC/ODBC server implemented here corresponds to the HiveServer2 in Hive 1.2.1 You can test the JDBC server with the beeline script that comes with either Spark or Hive 1.2.1.

To start the JDBC/ODBC server, run the following in the Spark directory:

./sbin/start-thriftserver.sh

This script accepts all bin/spark-submit command line options, plus a --hiveconf option to specify Hive properties. You may run ./sbin/start-thriftserver.sh --help for a complete list of all available options. By default, the server listens on localhost:10000. You may override this behaviour via either environment variables, i.e.:

export HIVE_SERVER2_THRIFT_PORT=<listening-port>
export HIVE_SERVER2_THRIFT_BIND_HOST=<listening-host>
./sbin/start-thriftserver.sh \
  --master <master-uri> \
  ...

or system properties:

./sbin/start-thriftserver.sh \
  --hiveconf hive.server2.thrift.port=<listening-port> \
  --hiveconf hive.server2.thrift.bind.host=<listening-host> \
  --master <master-uri>
  ...

Now you can use beeline to test the Thrift JDBC/ODBC server:

./bin/beeline

Connect to the JDBC/ODBC server in beeline with:

beeline> !connect jdbc:hive2://localhost:10000

Beeline will ask you for a username and password. In non-secure mode, simply enter the username on your machine and a blank password. For secure mode, please follow the instructions given in the beeline documentation.

Configuration of Hive is done by placing your hive-site.xml, core-site.xml and hdfs-site.xml files in conf/.

You may also use the beeline script that comes with Hive.

Thrift JDBC server also supports sending thrift RPC messages over HTTP transport. Use the following setting to enable HTTP mode as system property or in hive-site.xml file in conf/:

hive.server2.transport.mode - Set this to value: http
hive.server2.thrift.http.port - HTTP port number fo listen on; default is 10001
hive.server2.http.endpoint - HTTP endpoint; default is cliservice

To test, use beeline to connect to the JDBC/ODBC server in http mode with:

beeline> !connect jdbc:hive2://<host>:<port>/<database>?hive.server2.transport.mode=http;hive.server2.thrift.http.path=<http_endpoint>

Running the Spark SQL CLI

The Spark SQL CLI is a convenient tool to run the Hive metastore service in local mode and execute queries input from the command line. Note that the Spark SQL CLI cannot talk to the Thrift JDBC server.

To start the Spark SQL CLI, run the following in the Spark directory:

./bin/spark-sql

Configuration of Hive is done by placing your hive-site.xml, core-site.xml and hdfs-site.xml files in conf/. You may run ./bin/spark-sql --help for a complete list of all available options.

Migration Guide

Upgrading From Spark SQL 1.5 to 1.6

   ./sbin/start-thriftserver.sh \
     --conf spark.sql.hive.thriftServer.singleSession=true \
     ...
   

Upgrading From Spark SQL 1.4 to 1.5

Upgrading from Spark SQL 1.3 to 1.4

DataFrame data reader/writer interface

Based on user feedback, we created a new, more fluid API for reading data in (SQLContext.read) and writing data out (DataFrame.write), and deprecated the old APIs (e.g. SQLContext.parquetFile, SQLContext.jsonFile).

See the API docs for SQLContext.read ( Scala, Java, Python ) and DataFrame.write ( Scala, Java, Python ) more information.

DataFrame.groupBy retains grouping columns

Based on user feedback, we changed the default behavior of DataFrame.groupBy().agg() to retain the grouping columns in the resulting DataFrame. To keep the behavior in 1.3, set spark.sql.retainGroupColumns to false.

// In 1.3.x, in order for the grouping column "department" to show up,
// it must be included explicitly as part of the agg function call.
df.groupBy("department").agg($"department", max("age"), sum("expense"))

// In 1.4+, grouping column "department" is included automatically.
df.groupBy("department").agg(max("age"), sum("expense"))

// Revert to 1.3 behavior (not retaining grouping column) by:
sqlContext.setConf("spark.sql.retainGroupColumns", "false")
// In 1.3.x, in order for the grouping column "department" to show up,
// it must be included explicitly as part of the agg function call.
df.groupBy("department").agg(col("department"), max("age"), sum("expense"));

// In 1.4+, grouping column "department" is included automatically.
df.groupBy("department").agg(max("age"), sum("expense"));

// Revert to 1.3 behavior (not retaining grouping column) by:
sqlContext.setConf("spark.sql.retainGroupColumns", "false");
import pyspark.sql.functions as func

# In 1.3.x, in order for the grouping column "department" to show up,
# it must be included explicitly as part of the agg function call.
df.groupBy("department").agg("department"), func.max("age"), func.sum("expense"))

# In 1.4+, grouping column "department" is included automatically.
df.groupBy("department").agg(func.max("age"), func.sum("expense"))

# Revert to 1.3.x behavior (not retaining grouping column) by:
sqlContext.setConf("spark.sql.retainGroupColumns", "false")

Behavior change on DataFrame.withColumn

Prior to 1.4, DataFrame.withColumn() supports adding a column only. The column will always be added as a new column with its specified name in the result DataFrame even if there may be any existing columns of the same name. Since 1.4, DataFrame.withColumn() supports adding a column of a different name from names of all existing columns or replacing existing columns of the same name.

Note that this change is only for Scala API, not for PySpark and SparkR.

Upgrading from Spark SQL 1.0-1.2 to 1.3

In Spark 1.3 we removed the “Alpha” label from Spark SQL and as part of this did a cleanup of the available APIs. From Spark 1.3 onwards, Spark SQL will provide binary compatibility with other releases in the 1.X series. This compatibility guarantee excludes APIs that are explicitly marked as unstable (i.e., DeveloperAPI or Experimental).

Rename of SchemaRDD to DataFrame

The largest change that users will notice when upgrading to Spark SQL 1.3 is that SchemaRDD has been renamed to DataFrame. This is primarily because DataFrames no longer inherit from RDD directly, but instead provide most of the functionality that RDDs provide though their own implementation. DataFrames can still be converted to RDDs by calling the .rdd method.

In Scala there is a type alias from SchemaRDD to DataFrame to provide source compatibility for some use cases. It is still recommended that users update their code to use DataFrame instead. Java and Python users will need to update their code.

Unification of the Java and Scala APIs

Prior to Spark 1.3 there were separate Java compatible classes (JavaSQLContext and JavaSchemaRDD) that mirrored the Scala API. In Spark 1.3 the Java API and Scala API have been unified. Users of either language should use SQLContext and DataFrame. In general theses classes try to use types that are usable from both languages (i.e. Array instead of language specific collections). In some cases where no common type exists (e.g., for passing in closures or Maps) function overloading is used instead.

Additionally the Java specific types API has been removed. Users of both Scala and Java should use the classes present in org.apache.spark.sql.types to describe schema programmatically.

Isolation of Implicit Conversions and Removal of dsl Package (Scala-only)

Many of the code examples prior to Spark 1.3 started with import sqlContext._, which brought all of the functions from sqlContext into scope. In Spark 1.3 we have isolated the implicit conversions for converting RDDs into DataFrames into an object inside of the SQLContext. Users should now write import sqlContext.implicits._.

Additionally, the implicit conversions now only augment RDDs that are composed of Products (i.e., case classes or tuples) with a method toDF, instead of applying automatically.

When using function inside of the DSL (now replaced with the DataFrame API) users used to import org.apache.spark.sql.catalyst.dsl. Instead the public dataframe functions API should be used: import org.apache.spark.sql.functions._.

Removal of the type aliases in org.apache.spark.sql for DataType (Scala-only)

Spark 1.3 removes the type aliases that were present in the base sql package for DataType. Users should instead import the classes in org.apache.spark.sql.types

UDF Registration Moved to sqlContext.udf (Java & Scala)

Functions that are used to register UDFs, either for use in the DataFrame DSL or SQL, have been moved into the udf object in SQLContext.

sqlContext.udf.register("strLen", (s: String) => s.length())
sqlContext.udf().register("strLen", (String s) -> s.length(), DataTypes.IntegerType);

Python UDF registration is unchanged.

Python DataTypes No Longer Singletons

When using DataTypes in Python you will need to construct them (i.e. StringType()) instead of referencing a singleton.

Compatibility with Apache Hive

Spark SQL is designed to be compatible with the Hive Metastore, SerDes and UDFs. Currently Hive SerDes and UDFs are based on Hive 1.2.1, and Spark SQL can be connected to different versions of Hive Metastore (from 0.12.0 to 1.2.1. Also see [Interacting with Different Versions of Hive Metastore] (#interacting-with-different-versions-of-hive-metastore)).

Deploying in Existing Hive Warehouses

The Spark SQL Thrift JDBC server is designed to be “out of the box” compatible with existing Hive installations. You do not need to modify your existing Hive Metastore or change the data placement or partitioning of your tables.

Supported Hive Features

Spark SQL supports the vast majority of Hive features, such as:

Unsupported Hive Functionality

Below is a list of Hive features that we don’t support yet. Most of these features are rarely used in Hive deployments.

Major Hive Features

Esoteric Hive Features

Hive Input/Output Formats

Hive Optimizations

A handful of Hive optimizations are not yet included in Spark. Some of these (such as indexes) are less important due to Spark SQL’s in-memory computational model. Others are slotted for future releases of Spark SQL.

Reference

Data Types

Spark SQL and DataFrames support the following data types:

All data types of Spark SQL are located in the package org.apache.spark.sql.types. You can access them by doing

import  org.apache.spark.sql.types._
Data type Value type in Scala API to access or create a data type
ByteType Byte ByteType
ShortType Short ShortType
IntegerType Int IntegerType
LongType Long LongType
FloatType Float FloatType
DoubleType Double DoubleType
DecimalType java.math.BigDecimal DecimalType
StringType String StringType
BinaryType Array[Byte] BinaryType
BooleanType Boolean BooleanType
TimestampType java.sql.Timestamp TimestampType
DateType java.sql.Date DateType
ArrayType scala.collection.Seq ArrayType(elementType, [containsNull])
Note: The default value of containsNull is true.
MapType scala.collection.Map MapType(keyType, valueType, [valueContainsNull])
Note: The default value of valueContainsNull is true.
StructType org.apache.spark.sql.Row StructType(fields)
Note: fields is a Seq of StructFields. Also, two fields with the same name are not allowed.
StructField The value type in Scala of the data type of this field (For example, Int for a StructField with the data type IntegerType) StructField(name, dataType, nullable)

All data types of Spark SQL are located in the package of org.apache.spark.sql.types. To access or create a data type, please use factory methods provided in org.apache.spark.sql.types.DataTypes.

Data type Value type in Java API to access or create a data type
ByteType byte or Byte DataTypes.ByteType
ShortType short or Short DataTypes.ShortType
IntegerType int or Integer DataTypes.IntegerType
LongType long or Long DataTypes.LongType
FloatType float or Float DataTypes.FloatType
DoubleType double or Double DataTypes.DoubleType
DecimalType java.math.BigDecimal DataTypes.createDecimalType()
DataTypes.createDecimalType(precision, scale).
StringType String DataTypes.StringType
BinaryType byte[] DataTypes.BinaryType
BooleanType boolean or Boolean DataTypes.BooleanType
TimestampType java.sql.Timestamp DataTypes.TimestampType
DateType java.sql.Date DataTypes.DateType
ArrayType java.util.List DataTypes.createArrayType(elementType)
Note: The value of containsNull will be true
DataTypes.createArrayType(elementType, containsNull).
MapType java.util.Map DataTypes.createMapType(keyType, valueType)
Note: The value of valueContainsNull will be true.
DataTypes.createMapType(keyType, valueType, valueContainsNull)
StructType org.apache.spark.sql.Row DataTypes.createStructType(fields)
Note: fields is a List or an array of StructFields. Also, two fields with the same name are not allowed.
StructField The value type in Java of the data type of this field (For example, int for a StructField with the data type IntegerType) DataTypes.createStructField(name, dataType, nullable)

All data types of Spark SQL are located in the package of pyspark.sql.types. You can access them by doing

from pyspark.sql.types import *
Data type Value type in Python API to access or create a data type
ByteType int or long
Note: Numbers will be converted to 1-byte signed integer numbers at runtime. Please make sure that numbers are within the range of -128 to 127.
ByteType()
ShortType int or long
Note: Numbers will be converted to 2-byte signed integer numbers at runtime. Please make sure that numbers are within the range of -32768 to 32767.
ShortType()
IntegerType int or long IntegerType()
LongType long
Note: Numbers will be converted to 8-byte signed integer numbers at runtime. Please make sure that numbers are within the range of -9223372036854775808 to 9223372036854775807. Otherwise, please convert data to decimal.Decimal and use DecimalType.
LongType()
FloatType float
Note: Numbers will be converted to 4-byte single-precision floating point numbers at runtime.
FloatType()
DoubleType float DoubleType()
DecimalType decimal.Decimal DecimalType()
StringType string StringType()
BinaryType bytearray BinaryType()
BooleanType bool BooleanType()
TimestampType datetime.datetime TimestampType()
DateType datetime.date DateType()
ArrayType list, tuple, or array ArrayType(elementType, [containsNull])
Note: The default value of containsNull is True.
MapType dict MapType(keyType, valueType, [valueContainsNull])
Note: The default value of valueContainsNull is True.
StructType list or tuple StructType(fields)
Note: fields is a Seq of StructFields. Also, two fields with the same name are not allowed.
StructField The value type in Python of the data type of this field (For example, Int for a StructField with the data type IntegerType) StructField(name, dataType, nullable)
Data type Value type in R API to access or create a data type
ByteType integer
Note: Numbers will be converted to 1-byte signed integer numbers at runtime. Please make sure that numbers are within the range of -128 to 127.
"byte"
ShortType integer
Note: Numbers will be converted to 2-byte signed integer numbers at runtime. Please make sure that numbers are within the range of -32768 to 32767.
"short"
IntegerType integer "integer"
LongType integer
Note: Numbers will be converted to 8-byte signed integer numbers at runtime. Please make sure that numbers are within the range of -9223372036854775808 to 9223372036854775807. Otherwise, please convert data to decimal.Decimal and use DecimalType.
"long"
FloatType numeric
Note: Numbers will be converted to 4-byte single-precision floating point numbers at runtime.
"float"
DoubleType numeric "double"
DecimalType Not supported Not supported
StringType character "string"
BinaryType raw "binary"
BooleanType logical "bool"
TimestampType POSIXct "timestamp"
DateType Date "date"
ArrayType vector or list list(type="array", elementType=elementType, containsNull=[containsNull])
Note: The default value of containsNull is True.
MapType environment list(type="map", keyType=keyType, valueType=valueType, valueContainsNull=[valueContainsNull])
Note: The default value of valueContainsNull is True.
StructType named list list(type="struct", fields=fields)
Note: fields is a Seq of StructFields. Also, two fields with the same name are not allowed.
StructField The value type in R of the data type of this field (For example, integer for a StructField with the data type IntegerType) list(name=name, type=dataType, nullable=nullable)

NaN Semantics

There is specially handling for not-a-number (NaN) when dealing with float or double types that does not exactly match standard floating point semantics. Specifically: