Related
Suppose I have a dataframe such as the following:
import org.apache.spark.sql.{Row, DataFrame, SparkSession}
import org.apache.spark.sql.types.{StructType, StructField, IntegerType, StringType, DoubleType, NumericType}
import org.apache.spark.sql.functions.{udf, col, skewness}
val someData = Seq(
Row(8, "bat"),
Row(64, "mouse"),
Row(-27, "horse"),
Row(null, "mouse"),
Row(27, null)
)
val someSchema = List(
StructField("number", IntegerType, true),
StructField("word", StringType, true)
)
val someDF = spark.createDataFrame(
spark.sparkContext.parallelize(someData),
StructType(someSchema)
)
val df = someDF.withColumn("constantColumn", lit(1))
I would like to calculate the skewness of each column having a NumericType-like type. Then, if a column's skewness is above a certain threshold, I would like to transform it via f(x) = log(x + 1). (I know that performing log-transforms on negative data will give a NaN, but I would like to eventually write code that will take this possibility into account).
What I have tried so far:
I have found a way to do it, but it requires a mutable dataframe df. From my limited understanding, this is not desirable.
val log1p = scala.math.log1p(_)
val log1pUDF = udf(scala.math.log1p(_: Double))
val transformThreshold = 0.04
// filter those columns which have a type that inherits from NumericType
val numericColumns = df.columns.filter(column => df.select(column).schema(0).dataType.isInstanceOf[NumericType])
// for columns having NumericType, filter those that are sufficiently skewed
val columnsToTransform = numericColumns.filter(numericColumn => df.select(skewness(df(numericColumn))).head.getDouble(0) > transformThreshold)
// for all columns that are sufficiently skewed, perform log1p transform and add it to df
for(column <- columnsToTransform) {
// df should be mutable here!
df = df.withColumn(column + "_log1p", log1pUDF(df(column)))
}
My questions:
How can I achieve the goal without using mutable dataframes?
Is there an easier/quicker way to achieve what I have tried to do?
(Run on Spark 2.4.0, Scala 2.11.12.)
Instead of for() structure, you could use a recursive function :
def rec(df: DataFrame, columns: List[String]): DataFrame = columns match {
case Nil => df
case h :: xs => rec(df.withColumn(s"${h}_log1p", log1pUDF(col(h))), xs)
}
How can I convert an RDD (org.apache.spark.rdd.RDD[org.apache.spark.sql.Row]) to a Dataframe org.apache.spark.sql.DataFrame. I converted a dataframe to rdd using .rdd. After processing it I want it back in dataframe. How can I do this ?
This code works perfectly from Spark 2.x with Scala 2.11
Import necessary classes
import org.apache.spark.sql.{Row, SparkSession}
import org.apache.spark.sql.types.{DoubleType, StringType, StructField, StructType}
Create SparkSession Object, and Here it's spark
val spark: SparkSession = SparkSession.builder.master("local").getOrCreate
val sc = spark.sparkContext // Just used to create test RDDs
Let's an RDD to make it DataFrame
val rdd = sc.parallelize(
Seq(
("first", Array(2.0, 1.0, 2.1, 5.4)),
("test", Array(1.5, 0.5, 0.9, 3.7)),
("choose", Array(8.0, 2.9, 9.1, 2.5))
)
)
##Method 1
Using SparkSession.createDataFrame(RDD obj).
val dfWithoutSchema = spark.createDataFrame(rdd)
dfWithoutSchema.show()
+------+--------------------+
| _1| _2|
+------+--------------------+
| first|[2.0, 1.0, 2.1, 5.4]|
| test|[1.5, 0.5, 0.9, 3.7]|
|choose|[8.0, 2.9, 9.1, 2.5]|
+------+--------------------+
##Method 2
Using SparkSession.createDataFrame(RDD obj) and specifying column names.
val dfWithSchema = spark.createDataFrame(rdd).toDF("id", "vals")
dfWithSchema.show()
+------+--------------------+
| id| vals|
+------+--------------------+
| first|[2.0, 1.0, 2.1, 5.4]|
| test|[1.5, 0.5, 0.9, 3.7]|
|choose|[8.0, 2.9, 9.1, 2.5]|
+------+--------------------+
##Method 3 (Actual answer to the question)
This way requires the input rdd should be of type RDD[Row].
val rowsRdd: RDD[Row] = sc.parallelize(
Seq(
Row("first", 2.0, 7.0),
Row("second", 3.5, 2.5),
Row("third", 7.0, 5.9)
)
)
create the schema
val schema = new StructType()
.add(StructField("id", StringType, true))
.add(StructField("val1", DoubleType, true))
.add(StructField("val2", DoubleType, true))
Now apply both rowsRdd and schema to createDataFrame()
val df = spark.createDataFrame(rowsRdd, schema)
df.show()
+------+----+----+
| id|val1|val2|
+------+----+----+
| first| 2.0| 7.0|
|second| 3.5| 2.5|
| third| 7.0| 5.9|
+------+----+----+
SparkSession has a number of createDataFrame methods that create a DataFrame given an RDD. I imagine one of these will work for your context.
For example:
def createDataFrame(rowRDD: RDD[Row], schema: StructType): DataFrame
Creates a DataFrame from an RDD containing Rows using the given
schema.
Assuming your RDD[row] is called rdd, you can use:
val sqlContext = new SQLContext(sc)
import sqlContext.implicits._
rdd.toDF()
Note: This answer was originally posted here
I am posting this answer because I would like to share additional details about the available options that I did not find in the other answers
To create a DataFrame from an RDD of Rows, there are two main options:
1) As already pointed out, you could use toDF() which can be imported by import sqlContext.implicits._. However, this approach only works for the following types of RDDs:
RDD[Int]
RDD[Long]
RDD[String]
RDD[T <: scala.Product]
(source: Scaladoc of the SQLContext.implicits object)
The last signature actually means that it can work for an RDD of tuples or an RDD of case classes (because tuples and case classes are subclasses of scala.Product).
So, to use this approach for an RDD[Row], you have to map it to an RDD[T <: scala.Product]. This can be done by mapping each row to a custom case class or to a tuple, as in the following code snippets:
val df = rdd.map({
case Row(val1: String, ..., valN: Long) => (val1, ..., valN)
}).toDF("col1_name", ..., "colN_name")
or
case class MyClass(val1: String, ..., valN: Long = 0L)
val df = rdd.map({
case Row(val1: String, ..., valN: Long) => MyClass(val1, ..., valN)
}).toDF("col1_name", ..., "colN_name")
The main drawback of this approach (in my opinion) is that you have to explicitly set the schema of the resulting DataFrame in the map function, column by column. Maybe this can be done programatically if you don't know the schema in advance, but things can get a little messy there. So, alternatively, there is another option:
2) You can use createDataFrame(rowRDD: RDD[Row], schema: StructType) as in the accepted answer, which is available in the SQLContext object. Example for converting an RDD of an old DataFrame:
val rdd = oldDF.rdd
val newDF = oldDF.sqlContext.createDataFrame(rdd, oldDF.schema)
Note that there is no need to explicitly set any schema column. We reuse the old DF's schema, which is of StructType class and can be easily extended. However, this approach sometimes is not possible, and in some cases can be less efficient than the first one.
Suppose you have a DataFrame and you want to do some modification on the fields data by converting it to RDD[Row].
val aRdd = aDF.map(x=>Row(x.getAs[Long]("id"),x.getAs[List[String]]("role").head))
To convert back to DataFrame from RDD we need to define the structure type of the RDD.
If the datatype was Long then it will become as LongType in structure.
If String then StringType in structure.
val aStruct = new StructType(Array(StructField("id",LongType,nullable = true),StructField("role",StringType,nullable = true)))
Now you can convert the RDD to DataFrame using the createDataFrame method.
val aNamedDF = sqlContext.createDataFrame(aRdd,aStruct)
Method 1: (Scala)
val sqlContext = new org.apache.spark.sql.SQLContext(sc)
import sqlContext.implicits._
val df_2 = sc.parallelize(Seq((1L, 3.0, "a"), (2L, -1.0, "b"), (3L, 0.0, "c"))).toDF("x", "y", "z")
Method 2: (Scala)
case class temp(val1: String,val3 : Double)
val rdd = sc.parallelize(Seq(
Row("foo", 0.5), Row("bar", 0.0)
))
val rows = rdd.map({case Row(val1:String,val3:Double) => temp(val1,val3)}).toDF()
rows.show()
Method 1: (Python)
from pyspark.sql import Row
l = [('Alice',2)]
Person = Row('name','age')
rdd = sc.parallelize(l)
person = rdd.map(lambda r:Person(*r))
df2 = sqlContext.createDataFrame(person)
df2.show()
Method 2: (Python)
from pyspark.sql.types import *
l = [('Alice',2)]
rdd = sc.parallelize(l)
schema = StructType([StructField ("name" , StringType(), True) ,
StructField("age" , IntegerType(), True)])
df3 = sqlContext.createDataFrame(rdd, schema)
df3.show()
Extracted the value from the row object and then applied the case class to convert rdd to DF
val temp1 = attrib1.map{case Row ( key: Int ) => s"$key" }
val temp2 = attrib2.map{case Row ( key: Int) => s"$key" }
case class RLT (id: String, attrib_1 : String, attrib_2 : String)
import hiveContext.implicits._
val df = result.map{ s => RLT(s(0),s(1),s(2)) }.toDF
Here is a simple example of converting your List into Spark RDD and then converting that Spark RDD into Dataframe.
Please note that I have used Spark-shell's scala REPL to execute following code, Here sc is an instance of SparkContext which is implicitly available in Spark-shell. Hope it answer your question.
scala> val numList = List(1,2,3,4,5)
numList: List[Int] = List(1, 2, 3, 4, 5)
scala> val numRDD = sc.parallelize(numList)
numRDD: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[80] at parallelize at <console>:28
scala> val numDF = numRDD.toDF
numDF: org.apache.spark.sql.DataFrame = [_1: int]
scala> numDF.show
+---+
| _1|
+---+
| 1|
| 2|
| 3|
| 4|
| 5|
+---+
On newer versions of spark (2.0+)
import org.apache.spark.sql.SparkSession
import org.apache.spark.sql.functions._
import org.apache.spark.sql._
import org.apache.spark.sql.types._
val spark = SparkSession
.builder()
.getOrCreate()
import spark.implicits._
val dfSchema = Seq("col1", "col2", "col3")
rdd.toDF(dfSchema: _*)
One needs to create a schema, and attach it to the Rdd.
Assuming val spark is a product of a SparkSession.builder...
import org.apache.spark._
import org.apache.spark.sql._
import org.apache.spark.sql.types._
/* Lets gin up some sample data:
* As RDD's and dataframes can have columns of differing types, lets make our
* sample data a three wide, two tall, rectangle of mixed types.
* A column of Strings, a column of Longs, and a column of Doubules
*/
val arrayOfArrayOfAnys = Array.ofDim[Any](2,3)
arrayOfArrayOfAnys(0)(0)="aString"
arrayOfArrayOfAnys(0)(1)=0L
arrayOfArrayOfAnys(0)(2)=3.14159
arrayOfArrayOfAnys(1)(0)="bString"
arrayOfArrayOfAnys(1)(1)=9876543210L
arrayOfArrayOfAnys(1)(2)=2.71828
/* The way to convert an anything which looks rectangular,
* (Array[Array[String]] or Array[Array[Any]] or Array[Row], ... ) into an RDD is to
* throw it into sparkContext.parallelize.
* http://spark.apache.org/docs/latest/api/scala/index.html#org.apache.spark.SparkContext shows
* the parallelize definition as
* def parallelize[T](seq: Seq[T], numSlices: Int = defaultParallelism)
* so in our case our ArrayOfArrayOfAnys is treated as a sequence of ArraysOfAnys.
* Will leave the numSlices as the defaultParallelism, as I have no particular cause to change it.
*/
val rddOfArrayOfArrayOfAnys=spark.sparkContext.parallelize(arrayOfArrayOfAnys)
/* We'll be using the sqlContext.createDataFrame to add a schema our RDD.
* The RDD which goes into createDataFrame is an RDD[Row] which is not what we happen to have.
* To convert anything one tall and several wide into a Row, one can use Row.fromSeq(thatThing.toSeq)
* As we have an RDD[somethingWeDontWant], we can map each of the RDD rows into the desired Row type.
*/
val rddOfRows=rddOfArrayOfArrayOfAnys.map(f=>
Row.fromSeq(f.toSeq)
)
/* Now to construct our schema. This needs to be a StructType of 1 StructField per column in our dataframe.
* https://spark.apache.org/docs/latest/api/scala/index.html#org.apache.spark.sql.types.StructField shows the definition as
* case class StructField(name: String, dataType: DataType, nullable: Boolean = true, metadata: Metadata = Metadata.empty)
* Will leave the two default values in place for each of the columns:
* nullability as true,
* metadata as an empty Map[String,Any]
*
*/
val schema = StructType(
StructField("colOfStrings", StringType) ::
StructField("colOfLongs" , LongType ) ::
StructField("colOfDoubles", DoubleType) ::
Nil
)
val df=spark.sqlContext.createDataFrame(rddOfRows,schema)
/*
* +------------+----------+------------+
* |colOfStrings|colOfLongs|colOfDoubles|
* +------------+----------+------------+
* | aString| 0| 3.14159|
* | bString|9876543210| 2.71828|
* +------------+----------+------------+
*/
df.show
Same steps, but with fewer val declarations:
val arrayOfArrayOfAnys=Array(
Array("aString",0L ,3.14159),
Array("bString",9876543210L,2.71828)
)
val rddOfRows=spark.sparkContext.parallelize(arrayOfArrayOfAnys).map(f=>Row.fromSeq(f.toSeq))
/* If one knows the datatypes, for instance from JDBC queries as to RDBC column metadata:
* Consider constructing the schema from an Array[StructField]. This would allow looping over
* the columns, with a match statement applying the appropriate sql datatypes as the second
* StructField arguments.
*/
val sf=new Array[StructField](3)
sf(0)=StructField("colOfStrings",StringType)
sf(1)=StructField("colOfLongs" ,LongType )
sf(2)=StructField("colOfDoubles",DoubleType)
val df=spark.sqlContext.createDataFrame(rddOfRows,StructType(sf.toList))
df.show
I tried to explain the solution using the word count problem.
1. Read the file using sc
Produce word count
Methods to create DF
rdd.toDF method
rdd.toDF("word","count")
spark.createDataFrame(rdd,schema)
Read file using spark
val rdd=sc.textFile("D://cca175/data/")
Rdd to Dataframe
val df=sc.textFile("D://cca175/data/").toDF("t1")
df.show
Method 1
Create word count RDD to Dataframe
val df=rdd.flatMap(x=>x.split(" ")).map(x=>(x,1)).reduceByKey((x,y)=>(x+y)).toDF("word","count")
Method2
Create Dataframe from Rdd
val df=spark.createDataFrame(wordRdd)
# with header
val df=spark.createDataFrame(wordRdd).toDF("word","count") df.show
Method3
Define Schema
import org.apache.spark.sql.types._
val schema=new StructType().
add(StructField("word",StringType,true)).
add(StructField("count",StringType,true))
Create RowRDD
import org.apache.spark.sql.Row
val rowRdd=wordRdd.map(x=>(Row(x._1,x._2)))
Create DataFrame from RDD with schema
val df=spark.createDataFrame(rowRdd,schema)
df.show
I meet the same problem, and finally solve it. It's quite simple and easy.
You have to add this code import sc.implicits._, sc means SQLContext. add this code you will get rdd.toDF() method.
Transform your rdd[RawData] to rdd[YourCaseClass]. For example, you have a rdd type like this rdd[(String, Integer, Long)], you can create a Case Class YourCaseClass(name: String, age: Integer, timestamp: Long) and convert raw rdd to rdd with YourCaseClass type, then you get rdd[YourCaseClass]
save rdd[YourCaseClass] to hive table. yourRdd.toDF().write.format("parquet").mode(SaveMode.Overwrite).insertInto(yourHiveTableName) Use case class to represent rdd type, we can avoid naming each column field or StructType related schema.
To convert an Array[Row] to DataFrame or Dataset, the following works elegantly:
Say, schema is the StructType for the row,then
val rows: Array[Row]=...
implicit val encoder = RowEncoder.apply(schema)
import spark.implicits._
rows.toDS
I have a second question around CosineSimilarity / ColumnSimilarities in Spark 2.1. I'm kinda new to scala and all the Spark environment and this is not really clear to me:
How can I get back the ColumnSimilarities for each combination of columns from the rowMatrix in spark. Here is what I tried:
Data:
import org.apache.spark.sql.{SQLContext, Row, DataFrame}
import org.apache.spark.sql.types.{StructType, StructField, StringType, IntegerType, DoubleType}
import org.apache.spark.sql.functions._
// rdd
val rowsRdd: RDD[Row] = sc.parallelize(
Seq(
Row(2.0, 7.0, 1.0),
Row(3.5, 2.5, 0.0),
Row(7.0, 5.9, 0.0)
)
)
// Schema
val schema = new StructType()
.add(StructField("item_1", DoubleType, true))
.add(StructField("item_2", DoubleType, true))
.add(StructField("item_3", DoubleType, true))
// Data frame
val df = spark.createDataFrame(rowsRdd, schema)
Code:
import org.apache.spark.ml.feature.VectorAssembler
import org.apache.spark.mllib.linalg.distributed.{MatrixEntry, CoordinateMatrix, RowMatrix}
val rows = new VectorAssembler().setInputCols(df.columns).setOutputCol("vs")
.transform(df)
.select("vs")
.rdd
val items_mllib_vector = rows.map(_.getAs[org.apache.spark.ml.linalg.Vector](0))
.map(org.apache.spark.mllib.linalg.Vectors.fromML)
val mat = new RowMatrix(items_mllib_vector)
val simsPerfect = mat.columnSimilarities()
println("Pairwise similarities are: " + simsPerfect.entries.collect.mkString(", "))
Output:
Pairwise similarities are: MatrixEntry(0,2,0.24759378423606918), MatrixEntry(1,2,0.7376189553526812), MatrixEntry(0,1,0.8355316482961213)
So What I get is simsPerfect org.apache.spark.mllib.linalg.distributed.CoordinateMatrix of my Columns and similarities. How would I transform this back to a dataframe and get the right columns names with it?
My preferred output:
item_from | item_to | similarity
1 | 2 | 0.83 |
1 | 3 | 0.24 |
2 | 3 | 0.73 |
Thanks in advance
This approach also works without converting the row to String:
val transformedRDD = simsPerfect.entries.map{case MatrixEntry(row: Long, col:Long, sim:Double) => (row,col,sim)}
val dff = sqlContext.createDataFrame(transformedRDD).toDF("item_from", "item_to", "sim")
where, I assume val sqlContext = new org.apache.spark.sql.SQLContext(sc) is defined already and sc is the SparkContext.
I found a solution for my problem:
//Transform result to rdd
val transformedRDD = simsPerfect.entries.map{case MatrixEntry(row: Long, col:Long, sim:Double) => Array(row,col,sim).mkString(",")}
//Transform rdd[String] to rdd[Row]
val rdd2 = transformedRDD.map(a => Row(a))
// to DF
val dfschema = StructType(Array(StructField("value",StringType)))
val rddToDF = spark.createDataFrame(rdd2,dfschema)
//create new DF with schema
val newdf = rddToDF.select(expr("(split(value, ','))[0]").cast("string").as("item_from")
,expr("(split(value, ','))[1]").cast("string").as("item_to")
,expr("(split(value, ','))[2]").cast("string").as("sim"))
I'm sure there is another easier way to do this, but I'm happy that it works.
Acyally am working on spark 2.0.2
I would like to know , for example to work on logistic regression based on Spark ML.I would like to get each row of the dataframe into a vector which will be input for logistic regression , can you help get row resulted in the dataframe to get each row into a dense vector .Thanks.Here what i did to get the dataframe.
import org.apache.spark.ml.classification.LogisticRegression
import org.apache.spark.ml.linalg.{Vector, Vectors}
import org.apache.spark.ml.param.ParamMap
import org.apache.spark.sql.SparkSession
import org.apache.spark.sql.Row
import org.apache.hadoop.fs.shell.Display
object Example extends App {
val sparkSession = SparkSession.builder.master("local").appName("my-spark-app").getOrCreate()
val data=sparkSession.read.option("header", "true").csv("C://sample_lda_data.csv").toDF()
val data2=data.select("col2","col3","col4","col5","col6","col7","col8","col9")
at the end i want to get something like this as input for logistic regression
in the first position it will be the first column of the dataframe any help please
val data=sparkSession.read.option("header", "true").csv("C://sample_lda_data.csv").toDF()
val data2=data.select("col2","col3","col4","col5","col6","col7","col8","col9")
val assembler = new VectorAssembler().setInputCols(Array("col2", "col3", "col4")).setOutputCol("features")
val output = assembler.transform(data2)
main" java.lang.IllegalArgumentException: Data type StringType is not supported.
I'll be so gratefull.Thank you guys
You can use the array function and then map into LabeledPoints:
import org.apache.spark.mllib.linalg.Vectors
import org.apache.spark.mllib.regression.LabeledPoint
import org.apache.spark.sql._
import org.apache.spark.sql.functions._
import org.apache.spark.sql.types.DoubleType
// create an array column from all but first one:
val arrayCol: Column = array(df.columns.drop(1).map(col).map(_.cast(DoubleType)): _*)
// select array column and first column, and map into LabeledPoints
val result: Dataset[LabeledPoint] = df.select(col("col1").cast(DoubleType), arrayCol)
.map(r => LabeledPoint(
r.getAs[Double](0),
Vectors.dense(r.getAs[mutable.WrappedArray[Double]](1).toArray)
))
// You can use the Dataset or the RDD
result.show()
// +-----+---------------------+
// |label|features |
// +-----+---------------------+
// |1.0 |[2.0,3.0,4.0,0.5] |
// |11.0 |[12.0,13.0,14.0,15.0]|
// |21.0 |[22.0,23.0,24.0,25.0]|
// +-----+---------------------+
result.rdd.foreach(println)
// (1.0,[2.0,3.0,4.0,0.5])
// (21.0,[22.0,23.0,24.0,25.0])
I have wrote code to convert dataframe's numeric columns into dense vector. Please find below code. Note: here col1 and col2 are numeric type columns.
import sparksession.implicits._;
val result: Dataset[LabeledPoint] = df.map{ x => LabeledPoint(x.getAs[Integer]("Col1").toDouble, Vectors.dense(x.getAs[Double]("col2"))) }
result.show();
result.printSchema();
+-------+----------+
| label| features|
+-------+----------+
|31825.0| [75000.0]|
|58784.0| [24044.0]|
| 121.0| [41000.0]|
root
|-- label: double (nullable = true)
|-- features: vector (nullable = true)
How can I convert an RDD (org.apache.spark.rdd.RDD[org.apache.spark.sql.Row]) to a Dataframe org.apache.spark.sql.DataFrame. I converted a dataframe to rdd using .rdd. After processing it I want it back in dataframe. How can I do this ?
This code works perfectly from Spark 2.x with Scala 2.11
Import necessary classes
import org.apache.spark.sql.{Row, SparkSession}
import org.apache.spark.sql.types.{DoubleType, StringType, StructField, StructType}
Create SparkSession Object, and Here it's spark
val spark: SparkSession = SparkSession.builder.master("local").getOrCreate
val sc = spark.sparkContext // Just used to create test RDDs
Let's an RDD to make it DataFrame
val rdd = sc.parallelize(
Seq(
("first", Array(2.0, 1.0, 2.1, 5.4)),
("test", Array(1.5, 0.5, 0.9, 3.7)),
("choose", Array(8.0, 2.9, 9.1, 2.5))
)
)
##Method 1
Using SparkSession.createDataFrame(RDD obj).
val dfWithoutSchema = spark.createDataFrame(rdd)
dfWithoutSchema.show()
+------+--------------------+
| _1| _2|
+------+--------------------+
| first|[2.0, 1.0, 2.1, 5.4]|
| test|[1.5, 0.5, 0.9, 3.7]|
|choose|[8.0, 2.9, 9.1, 2.5]|
+------+--------------------+
##Method 2
Using SparkSession.createDataFrame(RDD obj) and specifying column names.
val dfWithSchema = spark.createDataFrame(rdd).toDF("id", "vals")
dfWithSchema.show()
+------+--------------------+
| id| vals|
+------+--------------------+
| first|[2.0, 1.0, 2.1, 5.4]|
| test|[1.5, 0.5, 0.9, 3.7]|
|choose|[8.0, 2.9, 9.1, 2.5]|
+------+--------------------+
##Method 3 (Actual answer to the question)
This way requires the input rdd should be of type RDD[Row].
val rowsRdd: RDD[Row] = sc.parallelize(
Seq(
Row("first", 2.0, 7.0),
Row("second", 3.5, 2.5),
Row("third", 7.0, 5.9)
)
)
create the schema
val schema = new StructType()
.add(StructField("id", StringType, true))
.add(StructField("val1", DoubleType, true))
.add(StructField("val2", DoubleType, true))
Now apply both rowsRdd and schema to createDataFrame()
val df = spark.createDataFrame(rowsRdd, schema)
df.show()
+------+----+----+
| id|val1|val2|
+------+----+----+
| first| 2.0| 7.0|
|second| 3.5| 2.5|
| third| 7.0| 5.9|
+------+----+----+
SparkSession has a number of createDataFrame methods that create a DataFrame given an RDD. I imagine one of these will work for your context.
For example:
def createDataFrame(rowRDD: RDD[Row], schema: StructType): DataFrame
Creates a DataFrame from an RDD containing Rows using the given
schema.
Assuming your RDD[row] is called rdd, you can use:
val sqlContext = new SQLContext(sc)
import sqlContext.implicits._
rdd.toDF()
Note: This answer was originally posted here
I am posting this answer because I would like to share additional details about the available options that I did not find in the other answers
To create a DataFrame from an RDD of Rows, there are two main options:
1) As already pointed out, you could use toDF() which can be imported by import sqlContext.implicits._. However, this approach only works for the following types of RDDs:
RDD[Int]
RDD[Long]
RDD[String]
RDD[T <: scala.Product]
(source: Scaladoc of the SQLContext.implicits object)
The last signature actually means that it can work for an RDD of tuples or an RDD of case classes (because tuples and case classes are subclasses of scala.Product).
So, to use this approach for an RDD[Row], you have to map it to an RDD[T <: scala.Product]. This can be done by mapping each row to a custom case class or to a tuple, as in the following code snippets:
val df = rdd.map({
case Row(val1: String, ..., valN: Long) => (val1, ..., valN)
}).toDF("col1_name", ..., "colN_name")
or
case class MyClass(val1: String, ..., valN: Long = 0L)
val df = rdd.map({
case Row(val1: String, ..., valN: Long) => MyClass(val1, ..., valN)
}).toDF("col1_name", ..., "colN_name")
The main drawback of this approach (in my opinion) is that you have to explicitly set the schema of the resulting DataFrame in the map function, column by column. Maybe this can be done programatically if you don't know the schema in advance, but things can get a little messy there. So, alternatively, there is another option:
2) You can use createDataFrame(rowRDD: RDD[Row], schema: StructType) as in the accepted answer, which is available in the SQLContext object. Example for converting an RDD of an old DataFrame:
val rdd = oldDF.rdd
val newDF = oldDF.sqlContext.createDataFrame(rdd, oldDF.schema)
Note that there is no need to explicitly set any schema column. We reuse the old DF's schema, which is of StructType class and can be easily extended. However, this approach sometimes is not possible, and in some cases can be less efficient than the first one.
Suppose you have a DataFrame and you want to do some modification on the fields data by converting it to RDD[Row].
val aRdd = aDF.map(x=>Row(x.getAs[Long]("id"),x.getAs[List[String]]("role").head))
To convert back to DataFrame from RDD we need to define the structure type of the RDD.
If the datatype was Long then it will become as LongType in structure.
If String then StringType in structure.
val aStruct = new StructType(Array(StructField("id",LongType,nullable = true),StructField("role",StringType,nullable = true)))
Now you can convert the RDD to DataFrame using the createDataFrame method.
val aNamedDF = sqlContext.createDataFrame(aRdd,aStruct)
Method 1: (Scala)
val sqlContext = new org.apache.spark.sql.SQLContext(sc)
import sqlContext.implicits._
val df_2 = sc.parallelize(Seq((1L, 3.0, "a"), (2L, -1.0, "b"), (3L, 0.0, "c"))).toDF("x", "y", "z")
Method 2: (Scala)
case class temp(val1: String,val3 : Double)
val rdd = sc.parallelize(Seq(
Row("foo", 0.5), Row("bar", 0.0)
))
val rows = rdd.map({case Row(val1:String,val3:Double) => temp(val1,val3)}).toDF()
rows.show()
Method 1: (Python)
from pyspark.sql import Row
l = [('Alice',2)]
Person = Row('name','age')
rdd = sc.parallelize(l)
person = rdd.map(lambda r:Person(*r))
df2 = sqlContext.createDataFrame(person)
df2.show()
Method 2: (Python)
from pyspark.sql.types import *
l = [('Alice',2)]
rdd = sc.parallelize(l)
schema = StructType([StructField ("name" , StringType(), True) ,
StructField("age" , IntegerType(), True)])
df3 = sqlContext.createDataFrame(rdd, schema)
df3.show()
Extracted the value from the row object and then applied the case class to convert rdd to DF
val temp1 = attrib1.map{case Row ( key: Int ) => s"$key" }
val temp2 = attrib2.map{case Row ( key: Int) => s"$key" }
case class RLT (id: String, attrib_1 : String, attrib_2 : String)
import hiveContext.implicits._
val df = result.map{ s => RLT(s(0),s(1),s(2)) }.toDF
Here is a simple example of converting your List into Spark RDD and then converting that Spark RDD into Dataframe.
Please note that I have used Spark-shell's scala REPL to execute following code, Here sc is an instance of SparkContext which is implicitly available in Spark-shell. Hope it answer your question.
scala> val numList = List(1,2,3,4,5)
numList: List[Int] = List(1, 2, 3, 4, 5)
scala> val numRDD = sc.parallelize(numList)
numRDD: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[80] at parallelize at <console>:28
scala> val numDF = numRDD.toDF
numDF: org.apache.spark.sql.DataFrame = [_1: int]
scala> numDF.show
+---+
| _1|
+---+
| 1|
| 2|
| 3|
| 4|
| 5|
+---+
On newer versions of spark (2.0+)
import org.apache.spark.sql.SparkSession
import org.apache.spark.sql.functions._
import org.apache.spark.sql._
import org.apache.spark.sql.types._
val spark = SparkSession
.builder()
.getOrCreate()
import spark.implicits._
val dfSchema = Seq("col1", "col2", "col3")
rdd.toDF(dfSchema: _*)
One needs to create a schema, and attach it to the Rdd.
Assuming val spark is a product of a SparkSession.builder...
import org.apache.spark._
import org.apache.spark.sql._
import org.apache.spark.sql.types._
/* Lets gin up some sample data:
* As RDD's and dataframes can have columns of differing types, lets make our
* sample data a three wide, two tall, rectangle of mixed types.
* A column of Strings, a column of Longs, and a column of Doubules
*/
val arrayOfArrayOfAnys = Array.ofDim[Any](2,3)
arrayOfArrayOfAnys(0)(0)="aString"
arrayOfArrayOfAnys(0)(1)=0L
arrayOfArrayOfAnys(0)(2)=3.14159
arrayOfArrayOfAnys(1)(0)="bString"
arrayOfArrayOfAnys(1)(1)=9876543210L
arrayOfArrayOfAnys(1)(2)=2.71828
/* The way to convert an anything which looks rectangular,
* (Array[Array[String]] or Array[Array[Any]] or Array[Row], ... ) into an RDD is to
* throw it into sparkContext.parallelize.
* http://spark.apache.org/docs/latest/api/scala/index.html#org.apache.spark.SparkContext shows
* the parallelize definition as
* def parallelize[T](seq: Seq[T], numSlices: Int = defaultParallelism)
* so in our case our ArrayOfArrayOfAnys is treated as a sequence of ArraysOfAnys.
* Will leave the numSlices as the defaultParallelism, as I have no particular cause to change it.
*/
val rddOfArrayOfArrayOfAnys=spark.sparkContext.parallelize(arrayOfArrayOfAnys)
/* We'll be using the sqlContext.createDataFrame to add a schema our RDD.
* The RDD which goes into createDataFrame is an RDD[Row] which is not what we happen to have.
* To convert anything one tall and several wide into a Row, one can use Row.fromSeq(thatThing.toSeq)
* As we have an RDD[somethingWeDontWant], we can map each of the RDD rows into the desired Row type.
*/
val rddOfRows=rddOfArrayOfArrayOfAnys.map(f=>
Row.fromSeq(f.toSeq)
)
/* Now to construct our schema. This needs to be a StructType of 1 StructField per column in our dataframe.
* https://spark.apache.org/docs/latest/api/scala/index.html#org.apache.spark.sql.types.StructField shows the definition as
* case class StructField(name: String, dataType: DataType, nullable: Boolean = true, metadata: Metadata = Metadata.empty)
* Will leave the two default values in place for each of the columns:
* nullability as true,
* metadata as an empty Map[String,Any]
*
*/
val schema = StructType(
StructField("colOfStrings", StringType) ::
StructField("colOfLongs" , LongType ) ::
StructField("colOfDoubles", DoubleType) ::
Nil
)
val df=spark.sqlContext.createDataFrame(rddOfRows,schema)
/*
* +------------+----------+------------+
* |colOfStrings|colOfLongs|colOfDoubles|
* +------------+----------+------------+
* | aString| 0| 3.14159|
* | bString|9876543210| 2.71828|
* +------------+----------+------------+
*/
df.show
Same steps, but with fewer val declarations:
val arrayOfArrayOfAnys=Array(
Array("aString",0L ,3.14159),
Array("bString",9876543210L,2.71828)
)
val rddOfRows=spark.sparkContext.parallelize(arrayOfArrayOfAnys).map(f=>Row.fromSeq(f.toSeq))
/* If one knows the datatypes, for instance from JDBC queries as to RDBC column metadata:
* Consider constructing the schema from an Array[StructField]. This would allow looping over
* the columns, with a match statement applying the appropriate sql datatypes as the second
* StructField arguments.
*/
val sf=new Array[StructField](3)
sf(0)=StructField("colOfStrings",StringType)
sf(1)=StructField("colOfLongs" ,LongType )
sf(2)=StructField("colOfDoubles",DoubleType)
val df=spark.sqlContext.createDataFrame(rddOfRows,StructType(sf.toList))
df.show
I tried to explain the solution using the word count problem.
1. Read the file using sc
Produce word count
Methods to create DF
rdd.toDF method
rdd.toDF("word","count")
spark.createDataFrame(rdd,schema)
Read file using spark
val rdd=sc.textFile("D://cca175/data/")
Rdd to Dataframe
val df=sc.textFile("D://cca175/data/").toDF("t1")
df.show
Method 1
Create word count RDD to Dataframe
val df=rdd.flatMap(x=>x.split(" ")).map(x=>(x,1)).reduceByKey((x,y)=>(x+y)).toDF("word","count")
Method2
Create Dataframe from Rdd
val df=spark.createDataFrame(wordRdd)
# with header
val df=spark.createDataFrame(wordRdd).toDF("word","count") df.show
Method3
Define Schema
import org.apache.spark.sql.types._
val schema=new StructType().
add(StructField("word",StringType,true)).
add(StructField("count",StringType,true))
Create RowRDD
import org.apache.spark.sql.Row
val rowRdd=wordRdd.map(x=>(Row(x._1,x._2)))
Create DataFrame from RDD with schema
val df=spark.createDataFrame(rowRdd,schema)
df.show
I meet the same problem, and finally solve it. It's quite simple and easy.
You have to add this code import sc.implicits._, sc means SQLContext. add this code you will get rdd.toDF() method.
Transform your rdd[RawData] to rdd[YourCaseClass]. For example, you have a rdd type like this rdd[(String, Integer, Long)], you can create a Case Class YourCaseClass(name: String, age: Integer, timestamp: Long) and convert raw rdd to rdd with YourCaseClass type, then you get rdd[YourCaseClass]
save rdd[YourCaseClass] to hive table. yourRdd.toDF().write.format("parquet").mode(SaveMode.Overwrite).insertInto(yourHiveTableName) Use case class to represent rdd type, we can avoid naming each column field or StructType related schema.
To convert an Array[Row] to DataFrame or Dataset, the following works elegantly:
Say, schema is the StructType for the row,then
val rows: Array[Row]=...
implicit val encoder = RowEncoder.apply(schema)
import spark.implicits._
rows.toDS