Related
I want to input many files and construct a pair(Array[String],Index) for each column, the index could be "file-i" where i is local column counter.
For example:
tableA.txt:00 01 02\n10 11 12
tableB.txt:03 04\n13 14
Target(each column with its filename and index):
RDD[Array[String],String] : (Array("00","10"),"tableA.txt-0"),(Array("01","11","tableA.txt-1"),(Array("02","12"),"tableA.txt-2"),(Array("03","13"),"tableB.txt-0"),(Array("04","14"),"tableB.txt-1")
My code:
val fc = classOf[TextInputFormat]
val kc = classOf[LongWritable]
val vc = classOf[Text]
val text = sc.newAPIHadoopFile(path, fc ,kc, vc, sc.hadoopConfiguration)
val linesWithFileNames = text.asInstanceOf[NewHadoopRDD[LongWritable, Text]]
.mapPartitionsWithInputSplit((inputSplit, iterator) => {
val file = inputSplit.asInstanceOf[FileSplit]
iterator.map(tup => (file.getPath, tup._2))
})
val columnsData = linesWithFileNames.flatMap(p => {
val filename = p._1.toString
val lines = p._2.toString.split("\n")
lines.map(l => l.split(" "))
.toSeq.transpose.zipWithIndex
.map(pair => (pair._1, filename+"-"+pair._2.toString))
})
My wrong result:
("00","tableA.txt-0"),("10","tableA.txt-0")...
One easy way to achieve what you want is to use wholeTextFiles which generates a RDD that associates each file path to its content.
The code would look like this:
val result : RDD[(Array[String], String)] = sc
.wholeTextFiles("data1")
.flatMap{ case (path, lines) => lines
.split("\\n")
.zipWithIndex
.map{ case (line, i) => (line.split("\\s+"),
path.split("/").last + "-" + i)}
}
Given a csv in the format below, what is the best way to load it into Scala as type Map[String, Array[String]], with the first key being the unique values for Col2, and the value Array[String]] as all co-occurring values of Col1?
a,1,
b,2,m
c,2,
d,1,
e,3,m
f,4,
g,2,
h,3,
I,1,
j,2,n
k,2,n
l,1,
m,5,
n,2,
I have tried to use the function below, but am getting errors trying to add to the Option type:
+= is not a member of Option[Array[String]]
In addition, I get overloaded method value ++ with alternatives:
with regards to the line case None => mapping ++ (linesplit(2) -> Array(linesplit(1)))
def parseCSV() : Map[String, Array[String]] = {
var mapping = Map[String, Array[String]]()
val lines = Source.fromFile("test.csv")
for (line <- lines.getLines) {
val linesplit = line.split(",")
mapping.get(linesplit(2)) match {
case Some(_) => mapping.get(linesplit(2)) += linesplit(1)
case None => mapping ++ (linesplit(2) -> Array(linesplit(1)))
}
}
mapping
}
}
I am hoping for a Map[String, Array[String]] like the following:
(2 -> Array["b","c","g","j", "k", "n"])
(3 -> Array["e","h"])
(4 -> Array["f"])
(5 -> Array["m"])
You can do the following:
First - read the file to List[List[String]]:
val rows: List[List[String]] = using(io.Source.fromFile("test.csv")) { source =>
source.getLines.toList map { line =>
line.split(",").map(_.trim).toList
}
}
Then, because the input has only 2 values per row, I filter the rows (rows with only one value I want to ignore)
val filteredRows = rows.filter(row => row.size > 1)
And the last step is to groupBy the first value (which is the second column - the index column is not returned from Source.fromFile):
filteredRows.groupBy(row => row.head).mapValues(_.map(_.last)))
This isn't complete, but it should give you an outline of how it might be done.
io.Source
.fromFile("so.txt") //open file
.getLines() //line by line
.map(_.split(",")) //split on commas
.toArray //load into memory
.groupMap(_(1))(_(0)) //Scala 2.13
//res0: Map[String,Array[String]] = Map(4 -> Array(f), 5 -> Array(m), 1 -> Array(a, d, I, l), 2 -> Array(b, c, g, j, k, n), 3 -> Array(e, h))
You'll notice that the file resource isn't closed, and it doesn't handle malformed input. I leave that for the diligent reader.
For the above code mutable Map & ArrayBuffer should be used, as they could be mutated/updated later.
def parseCSV(): Map[String, Array[String]] = {
val mapping = scala.collection.mutable.Map[String, ArrayBuffer[String]]()
val lines = Source.fromFile("test.csv")
for (line <- lines.getLines) {
val linesplit = line.split(",")
val key = line.split(",")(1)
val values = line.replace(s",$key", "").split(",")
mapping.get(key) match {
case Some(_) => mapping(linesplit(1)) ++= values
case None =>
val ab = ArrayBuffer[String]()
mapping(linesplit(1)) = ab ++= values
}
}
mapping.map(v => (v._1, v._2.toArray)).toMap
}
I have an rdd say sample_rdd of type RDD[(String, String, Int))] with 3 columns id,item,count. sample data:
id1|item1|1
id1|item2|3
id1|item3|4
id2|item1|3
id2|item4|2
I want to join each id against a lookup_rdd this:
item1|0
item2|0
item3|0
item4|0
item5|0
The output should give me following for id1, outerjoin with lookuptable:
item1|1
item2|3
item3|4
item4|0
item5|0
Similarly for id2 i should get:
item1|3
item2|0
item3|0
item4|2
item5|0
Finally output for each id should have all counts with id:
id1,1,3,4,0,0
id2,3,0,0,2,0
IMPORTANT:this output should be always ordered according to the order in lookup
This is what i have tried:
val line = rdd_sample.map { case (id, item, count) => (id, (item,count)) }.map(row=>(row._1,row._2)).groupByKey()
get(line).map(l=>(l._1,l._2)).mapValues(item_count=>lookup_rdd.leftOuterJoin(item_count))
def get (line: RDD[(String, Iterable[(String, Int)])]) = { for{ (id, item_cnt) <- line i = item_cnt.map(tuple => (tuple._1,tuple._2)) } yield (id,i)
Try below. Run each step on your local console to understand whats happening in detail.
The idea is to zipwithindex and form seq based on lookup_rdd.
(i1,0),(i2,1)..(i5,4) and (id1,0),(id2,1)
Index of final result wanted = [delta(length of lookup_rdd seq) * index of id1..id2 ] + index of i1...i5
So the base seq generated will be (0,(i1,id1)),(1,(i2,id1))...(8,(i4,id2)),(9,(i5,id2))
and then based on the key(i1,id1) reduce and calculate count.
val res2 = sc.parallelize(arr) //sample_rdd
val res3 = sc.parallelize(cart) //lookup_rdd
val delta = res3.count
val res83 = res3.map(_._1).zipWithIndex.cartesian(res2.map(_._1).distinct.zipWithIndex).map(x => (((x._1._1,x._2._1),((delta * x._2._2) + x._1._2, 0)))
val res86 = res2.map(x => ((x._2,x._1),x._3)).reduceByKey(_+_)
val res88 = res83.leftOuterJoin(res86)
val res91 = res88.map( x => {
x._2._2 match {
case Some(x1) => (x._2._1._1, (x._1,x._2._1._2+x1))
case None => (x._2._1._1, (x._1,x._2._1._2))
}
})
val res97 = res91.sortByKey(true).map( x => {
(x._2._1._2,List(x._2._2))}).reduceByKey(_++_)
res97.collect
// SOLUTION: Array((id1,List(1,3,4,0,0)),(id2,List(3,0,0,2,0)))
We are trying to generate column wise statistics of our dataset in spark. In addition to using the summary function from statistics library. We are using the following procedure:
We determine the columns with string values
Generate key value pair for the whole dataset, using the column number as key and value of column as value
generate a new map of format
(K,V) ->((K,V),1)
Then we use reduceByKey to find the sum of all unique value in all the columns. We cache this output to reduce further computation time.
In the next step we cycle through the columns using a for loop to find the statistics for all the columns.
We are trying to reduce the for loop by again utilizing the map reduce way but we are unable to find some way to achieve it. Doing so will allow us to generate column statistics for all columns in one execution. The for loop method is running sequentially making it very slow.
Code:
//drops the header
def dropHeader(data: RDD[String]): RDD[String] = {
data.mapPartitionsWithIndex((idx, lines) => {
if (idx == 0) {
lines.drop(1)
}
lines
})
}
def retAtrTuple(x: String) = {
val newX = x.split(",")
for (h <- 0 until newX.length)
yield (h,newX(h))
}
val line = sc.textFile("hdfs://.../myfile.csv")
val withoutHeader: RDD[String] = dropHeader(line)
val kvPairs = withoutHeader.flatMap(retAtrTuple) //generates a key-value pair where key is the column number and value is column's value
var bool_numeric_col = kvPairs.map{case (x,y) => (x,isNumeric(y))}.reduceByKey(_&&_).sortByKey() //this contains column indexes as key and boolean as value (true for numeric and false for string type)
var str_cols = bool_numeric_col.filter{case (x,y) => y == false}.map{case (x,y) => x}
var num_cols = bool_numeric_col.filter{case (x,y) => y == true}.map{case (x,y) => x}
var str_col = str_cols.toArray //array consisting the string col
var num_col = num_cols.toArray //array consisting numeric col
val colCount = kvPairs.map((_,1)).reduceByKey(_+_)
val e1 = colCount.map{case ((x,y),z) => (x,(y,z))}
var numPairs = e1.filter{case (x,(y,z)) => str_col.contains(x) }
//running for loops which needs to be parallelized/optimized as it sequentially operates on each column. Idea is to find the top10, bottom10 and number of distinct elements column wise
for(i <- str_col){
var total = numPairs.filter{case (x,(y,z)) => x==i}.sortBy(_._2._2)
var leastOnes = total.take(10)
println("leastOnes for Col" + i)
leastOnes.foreach(println)
var maxOnes = total.sortBy(-_._2._2).take(10)
println("maxOnes for Col" + i)
maxOnes.foreach(println)
println("distinct for Col" + i + " is " + total.count)
}
Let me simplify your question a bit. (A lot actually.) We have an RDD[(Int, String)] and we want to find the top 10 most common Strings for each Int (which are all in the 0–100 range).
Instead of sorting, as in your example, it is more efficient to use the Spark built-in RDD.top(n) method. Its run-time is linear in the size of the data, and requires moving much less data around than a sort.
Consider the implementation of top in RDD.scala. You want to do the same, but with one priority queue (heap) per Int key. The code becomes fairly complex:
import org.apache.spark.util.BoundedPriorityQueue // Pretend it's not private.
def top(n: Int, rdd: RDD[(Int, String)]): Map[Int, Iterable[String]] = {
// A heap that only keeps the top N values, so it has bounded size.
type Heap = BoundedPriorityQueue[(Long, String)]
// Get the word counts.
val counts: RDD[[(Int, String), Long)] =
rdd.map(_ -> 1L).reduceByKey(_ + _)
// In each partition create a column -> heap map.
val perPartition: RDD[Map[Int, Heap]] =
counts.mapPartitions { items =>
val heaps =
collection.mutable.Map[Int, Heap].withDefault(i => new Heap(n))
for (((k, v), count) <- items) {
heaps(k) += count -> v
}
Iterator.single(heaps)
}
// Merge the per-partition heap maps into one.
val merged: Map[Int, Heap] =
perPartition.reduce { (heaps1, heaps2) =>
val heaps =
collection.mutable.Map[Int, Heap].withDefault(i => new Heap(n))
for ((k, heap) <- heaps1.toSeq ++ heaps2.toSeq) {
for (cv <- heap) {
heaps(k) += cv
}
}
heaps
}
// Discard counts, return just the top strings.
merged.mapValues(_.map { case(count, value) => value })
}
This is efficient, but made painful because we need to work with multiple columns at the same time. It would be way easier to have one RDD per column and just call rdd.top(10) on each.
Unfortunately the naive way to split up the RDD into N smaller RDDs does N passes:
def split(together: RDD[(Int, String)], columns: Int): Seq[RDD[String]] = {
together.cache // We will make N passes over this RDD.
(0 until columns).map {
i => together.filter { case (key, value) => key == i }.values
}
}
A more efficient solution could be to write out the data into separate files by key, then load it back into separate RDDs. This is discussed in Write to multiple outputs by key Spark - one Spark job.
Thanks for #Daniel Darabos's answer. But there are some mistakes.
mixed use of Map and collection.mutable.Map
withDefault((i: Int) => new Heap(n)) do not create a new Heap when you set heaps(k) += count -> v
mix uasage of parentheses
Here is the modified code:
//import org.apache.spark.util.BoundedPriorityQueue // Pretend it's not private. copy to your own folder and import it
import org.apache.log4j.{Level, Logger}
import org.apache.spark.rdd.RDD
import org.apache.spark.{SparkConf, SparkContext}
object BoundedPriorityQueueTest {
// https://stackoverflow.com/questions/28166190/spark-column-wise-word-count
def top(n: Int, rdd: RDD[(Int, String)]): Map[Int, Iterable[String]] = {
// A heap that only keeps the top N values, so it has bounded size.
type Heap = BoundedPriorityQueue[(Long, String)]
// Get the word counts.
val counts: RDD[((Int, String), Long)] =
rdd.map(_ -> 1L).reduceByKey(_ + _)
// In each partition create a column -> heap map.
val perPartition: RDD[collection.mutable.Map[Int, Heap]] =
counts.mapPartitions { items =>
val heaps =
collection.mutable.Map[Int, Heap]() // .withDefault((i: Int) => new Heap(n))
for (((k, v), count) <- items) {
println("\n---")
println("before add " + ((k, v), count) + ", the map is: ")
println(heaps)
if (!heaps.contains(k)) {
println("not contains key " + k)
heaps(k) = new Heap(n)
println(heaps)
}
heaps(k) += count -> v
println("after add " + ((k, v), count) + ", the map is: ")
println(heaps)
}
println(heaps)
Iterator.single(heaps)
}
// Merge the per-partition heap maps into one.
val merged: collection.mutable.Map[Int, Heap] =
perPartition.reduce { (heaps1, heaps2) =>
val heaps =
collection.mutable.Map[Int, Heap]() //.withDefault((i: Int) => new Heap(n))
println(heaps)
for ((k, heap) <- heaps1.toSeq ++ heaps2.toSeq) {
for (cv <- heap) {
heaps(k) += cv
}
}
heaps
}
// Discard counts, return just the top strings.
merged.mapValues(_.map { case (count, value) => value }).toMap
}
def main(args: Array[String]): Unit = {
Logger.getRootLogger().setLevel(Level.FATAL) //http://stackoverflow.com/questions/27781187/how-to-stop-messages-displaying-on-spark-console
val conf = new SparkConf().setAppName("word count").setMaster("local[1]")
val sc = new SparkContext(conf)
sc.setLogLevel("WARN") //http://stackoverflow.com/questions/27781187/how-to-stop-messages-displaying-on-spark-console
val words = sc.parallelize(List((1, "s11"), (1, "s11"), (1, "s12"), (1, "s13"), (2, "s21"), (2, "s22"), (2, "s22"), (2, "s23")))
println("# words：" + words.count())
val result = top(1, words)
println("\n--result:")
println(result)
sc.stop()
print("DONE")
}
}
I'm trying to join two datasets based on two columns. It works until I use one column but fails with below error
:29: error: value join is not a member of org.apache.spark.rdd.RDD[(String, String, (String, String, String, String, Double))]
val finalFact = fact.join(dimensionWithSK).map { case(nk1,nk2, ((parts1,parts2,parts3,parts4,amount), (sk, prop1,prop2,prop3,prop4))) => (sk,amount) }
Code :
import org.apache.spark.rdd.RDD
def zipWithIndex[T](rdd: RDD[T]) = {
val partitionSizes = rdd.mapPartitions(p => Iterator(p.length)).collect
val ranges = partitionSizes.foldLeft(List((0, 0))) { case(accList, count) =>
val start = accList.head._2
val end = start + count
(start, end) :: accList
}.reverse.tail.toArray
rdd.mapPartitionsWithIndex( (index, partition) => {
val start = ranges(index)._1
val end = ranges(index)._2
val indexes = Iterator.range(start, end)
partition.zip(indexes)
})
}
val dimension = sc.
textFile("dimension.txt").
map{ line =>
val parts = line.split("\t")
(parts(0),parts(1),parts(2),parts(3),parts(4),parts(5))
}
val dimensionWithSK =
zipWithIndex(dimension).map { case((nk1,nk2,prop3,prop4,prop5,prop6), idx) => (nk1,nk2,(prop3,prop4,prop5,prop6,idx + nextSurrogateKey)) }
val fact = sc.
textFile("fact.txt").
map { line =>
val parts = line.split("\t")
// we need to output (Naturalkey, (FactId, Amount)) in
// order to be able to join with the dimension data.
(parts(0),parts(1), (parts(2),parts(3), parts(4),parts(5),parts(6).toDouble))
}
val finalFact = fact.join(dimensionWithSK).map { case(nk1,nk2, ((parts1,parts2,parts3,parts4,amount), (sk, prop1,prop2,prop3,prop4))) => (sk,amount) }
Request someone's help here..
Thanks
Sridhar
If you look at the signature of join it works on an RDD of pairs:
def join[W](other: RDD[(K, W)], partitioner: Partitioner): RDD[(K, (V, W))]
You have a triple. I guess your trying to join on the first 2 elements of the tuple, and so you need to map your triple to a pair, where the first element of the pair is a pair containing the first two elements of the triple, e.g. for any Types V1 and V2
val left: RDD[(String, String, V1)] = ??? // some rdd
val right: RDD[(String, String, V2)] = ??? // some rdd
left.map {
case (key1, key2, value) => ((key1, key2), value)
}
.join(
right.map {
case (key1, key2, value) => ((key1, key2), value)
})
This will give you an RDD of the form RDD[(String, String), (V1, V2)]
rdd1 Schema :
field1,field2, field3, fieldX,.....
rdd2 Schema :
field1, field2, field3, fieldY,.....
val joinResult = rdd1.join(rdd2,
Seq("field1", "field2", "field3"), "outer")
joinResult schema :
field1, field2, field3, fieldX, fieldY, ......
val emp = sc.
textFile("emp.txt").
map { line =>
val parts = line.split("\t")
// we need to output (Naturalkey, (FactId, Amount)) in
// order to be able to join with the dimension data.
((parts(0), parts(2)),parts(1))
}
val emp_new = sc.
textFile("emp_new.txt").
map { line =>
val parts = line.split("\t")
// we need to output (Naturalkey, (FactId, Amount)) in
// order to be able to join with the dimension data.
((parts(0), parts(2)),parts(1))
}
val finalemp =
emp_new.join(emp).
map { case((nk1,nk2) ,((parts1), (val1))) => (nk1,parts1,val1) }