// JOINS =====================================

Using RDD

//emp table

id,name,age

1,Vinayak, 35

2,Nilesh, 37

3,Raju, 30

4,Karthik, 28

5,Shreshta,1

6,Siddhish, 2

val emp = sc.textFile("C:/Users/Redirection/pateln7/Downloads/SPARK/emp.txt")

val first = emp.first

val empfile = emp.filter(x=>x!=first)

val empmap=empfile.map(x=>{

val splitted=x.split(",")

val id=splitted(0).trim.toInt

val name=splitted(1).trim

val age=splitted(2).trim.toInt

(id,(name,age))

})

empmap.foreach(println)

val output = empmap.filter(x=>x._2._2>30)

//payroll table

id,dept,sal

1,dev, 3500

2,dev, 3700

3,qa, 3000

4,qa, 2800

5,prod,1000

6,prod, 2000

val payroll = sc.textFile("C:/Users/Redirection/pateln7/Downloads/SPARK/payroll.txt")

val first = payroll.first

val payrollfile = payroll.filter(x=>x!=first)

val payrollmap=payrollfile.map(x=>{

val splitted=x.split(",")

val id=splitted(0).trim.toInt

val dept=splitted(1).trim

val sal=splitted(2).trim.toInt

(id,(dept,sal))

})

payrollmap.foreach(println)

val output = payrollmap.filter(x=>x._2._2>2000)

(1,(dev,3500))

(2,(dev,3700))

(4,(qa,2800))

(3,(qa,3000))

val joined=empmap.join(payrollmap)

joined.foreach(println)

(1,((Vinayak,35),(dev,3500)))

(3,((Raju,30),(qa,3000)))

(5,((Shreshta,1),(prod,1000)))

(4,((Karthik,28),(qa,2800)))

(6,((Siddhish,2),(prod,2000)))

(2,((Nilesh,37),(dev,3700)))

age > 30 and sal > 2000

val output=joined.filter(x=>x._2._1._2>30 && x._2._2._2>2000)

Using Dataframe

id,name,age

1,Vinayak, 35

2,Nilesh, 37

3,Raju, 30

4,Karthik, 28

5,Shreshta,1

6,Siddhish, 2

val emp = sc.textFile("C:/Users/Redirection/pateln7/Downloads/SPARK/emp.txt")

val first = emp.first

val empfile = emp.filter(x=>x!=first)

val empmap=empfile.map(x=>{

val splitted=x.split(",")

val id=splitted(0).trim.toInt

val name=splitted(1).trim

val age=splitted(2).trim.toInt

(id,name,age)

})

empmap.foreach(println)

val empdf = empmap.toDF("id","name","age")

//payroll table

id,dept,sal

1,dev, 3500

2,dev, 3700

3,qa, 3000

4,qa, 2800

5,prod,1000

6,prod, 2000

val payroll = sc.textFile("C:/Users/Redirection/pateln7/Downloads/SPARK/payroll.txt")

val first = payroll.first

val payrollfile = payroll.filter(x=>x!=first)

val payrollmap=payrollfile.map(x=>{

val splitted=x.split(",")

val id=splitted(0).trim.toInt

val dept=splitted(1).trim

val sal=splitted(2).trim.toInt

(id,dept,sal)

})

payrollmap.foreach(println)

val payrolldf = payrollmap.toDF("id","dept","sal")

scala> val dfjoined = empdf.join(payrolldf, empdf("id")===payrolldf("id"))

dfjoined: org.apache.spark.sql.DataFrame = [id: int, name: string ... 4 more fields]

scala> dfjoined.show

+---+--------+---+---+----+----+

| id| name|age| id|dept| sal|

+---+--------+---+---+----+----+

| 1| Vinayak| 35| 1| dev|3500|

| 6|Siddhish| 2| 6|prod|2000|

| 3| Raju| 30| 3| qa|3000|

| 5|Shreshta| 1| 5|prod|1000|

| 4| Karthik| 28| 4| qa|2800|

| 2| Nilesh| 37| 2| dev|3700|

+---+--------+---+---+----+----+

Using sqlcontext

val dfjoined = empdf.join(payrolldf, empdf("id")===payrolldf("id"))

dfjoined.registerTempTable("joinedemp")

val sqlContext = dfjoined.sqlContext

val output = sqlContext.sql("select * from joinedemp where age>30 and sal> 3500").show

Nächster

Vorher

Setup in windows