总述
Before Spark 2.0, the main programming interface of Spark was the Resilient Distributed Dataset (RDD).
After Spark 2.0, RDDs are replaced by Dataset, which is strongly-typed like an RDD, but with richer optimizations under the hood（底层）. The RDD interface is still supported. However, we highly recommend you to switch to use Dataset, which has better performance than RDD.
Datasets and DataFrames
- A Dataset is a distributed collection of data. Dataset is a new interface added in Spark 1.6 that provides the benefits of RDDs (strong typing, ability to use powerful lambda functions) with the benefits of Spark SQL’s optimized execution engine.
- A DataFrame is a Dataset 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.
create
./bin/spark-shell
scala> val textFile = spark.read.textFile("README.md")
textFile: org.apache.spark.sql.Dataset[String] = [value: string]
Caching
Spark also supports pulling data sets into a cluster-wide in-memory cache. This is very useful when data is accessed repeatedly, such as when querying a small “hot” dataset or when running an iterative algorithm like PageRank. As a simple example, let’s mark our linesWithSpark dataset to be cached:
scala> linesWithSpark.cache()
res7: linesWithSpark.type = [value: string]
scala> linesWithSpark.count()
res8: Long = 15
Interoperating（交互） with RDDs
Spark SQL supports two different methods for converting existing RDDs into Datasets. 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.
case class Person(name: String, age: Long)
// For implicit conversions from RDDs to DataFrames
import spark.implicits._
// Create an RDD of Person objects from a text file, convert it to a Dataframe
val peopleDF = spark.sparkContext
.textFile("examples/src/main/resources/people.txt")
.map(_.split(","))
.map(attributes => Person(attributes(0), attributes(1).trim.toInt))
.toDF()
// Register the DataFrame as a temporary view
peopleDF.createOrReplaceTempView("people")
The second method for creating Datasets 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 Datasets when the columns and their types are not known until runtime.
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)
import org.apache.spark.sql.types._
// Create an RDD
val peopleRDD = spark.sparkContext.textFile("examples/src/main/resources/people.txt")
// The schema is encoded in a string
val schemaString = "name age"
// Generate the schema based on the string of schema
val fields = schemaString.split(" ")
.map(fieldName => StructField(fieldName, StringType, nullable = true))
val schema = StructType(fields)
// Convert records of the RDD (people) to Rows
val rowRDD = peopleRDD
.map(_.split(","))
.map(attributes => Row(attributes(0), attributes(1).trim))
// Apply the schema to the RDD
val peopleDF = spark.createDataFrame(rowRDD, schema)
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. All built-in file sources (including Text/CSV/JSON/ORC/Parquet) are able to discover and infer partitioning information automatically. For example, with two extra columns, gender and country as partitioning columns:
data.parquet 里面仅有 两个字段 name，age
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 SparkSession.read.parquet or SparkSession.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)
spark.sql.sources.partitionColumnTypeInference.enabled
Notice that the data types of the partitioning columns are automatically inferred. Currently, numeric data types, date, timestamp 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.
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.
data source option mergeSchema OR spark.sql.parquet.mergeSchema
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
setting data source option mergeSchema to true when reading Parquet files (as shown in the examples below), or
setting the global SQL option spark.sql.parquet.mergeSchema to true.
sample
// This is used to implicitly convert an RDD to a DataFrame.
import spark.implicits._
// Create a simple DataFrame, store into a partition directory
val squaresDF = spark.sparkContext.makeRDD(1 to 5).map(i => (i, i * i)).toDF("value", "square")
squaresDF.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 cubesDF = spark.sparkContext.makeRDD(6 to 10).map(i => (i, i * i * i)).toDF("value", "cube")
cubesDF.write.parquet("data/test_table/key=2")
// Read the partitioned table
val mergedDF = spark.read.option("mergeSchema", "true").parquet("data/test_table")
mergedDF.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
// |-- value: int (nullable = true)
// |-- square: int (nullable = true)
// |-- cube: int (nullable = true)
// |-- key: int (nullable = true)