How to get result using better algorithm? - sql

Question

We have the list like below
List:
(("herry","0,1,2"),("herry","1,3"),("herry","3,6"),("herry","4"),("John","5"))
As the number in the string may be referred by different elements, the expected result is:
("herry","0,1,2,3,6"), ("herry","4"), ("John","5")
I worked out solution using scala, but it looks complicated, is there a more clean and easy way to work out the result? Thanks in advance!
Here is my solution in scala,
val foo=List(("herry","0,1,2"),("herry","1,3"),("herry","3,6"),("herry","4"),("John","5"))
println(GetValue)
def GetValue()={
foo.zipWithIndex.map((tuple: ((String, String), Int)) =>{
val tuples = getrelated(tuple._1, foo)
(tuple._2, tuples)
}).map((tuple: (Int, List[(String, String)])) => tuple._2)
.map((tuples: List[(String, String)]) => (tuples.head._1,tuples.map((tuple: (String, String)) => tuple._2)))
.map((tuple: (String, List[String])) => (tuple._1, tuple._2.mkString(",").split(",").distinct.sorted.mkString(",")))
.distinct
}
def getrelated(start:(String,String),fooList:List[(String,String)]):List[(String,String)]={
val fooListWithout = fooList.filter((tuple: (String, String)) => tuple != start)
val result=fooListWithout
.filter((tuple: (String, String)) => findmatching(tuple._2,start._2))
.flatMap((tuple: (String, String)) => start :: getrelated(tuple,fooListWithout))
if (result.isEmpty)
List(start)
else
result
}
def findmatching(key1:String,key2:String)={
(key1.split(",")++key2.split(","))
.groupBy(identity)
.mapValues((strings: Array[String]) => strings.size)
.exists((tuple: (String, Int)) => tuple._2>1)
}
Let me clarify the algorithm
if the number list has the overlapping number, then group these number as one element
if the number list has no overlapping number, then consider it as independent element
for example,
input: List(("herry","0,1,2"),("herry","1,3"),("herry","7,4"),("herry","4"),("John","5"))
expected output: List(("herry","0,1,2,3" ), ("herry","4,7"), ("John","5"))
input: List(("herry","0,1,2"),("herry","1,3"),("herry","3,6"),("herry","4"),("John","5"))
expected: List("herry","0,1,2,3,6"), ("herry","4"), ("John","5")
input: List(("herry","0,1"),("herry","2,3"),("herry","3,6"),("herry","4"),("John","5"))
expected: List("herry","0,1"),("herry","2,3,6"), ("herry","4"), ("John","5")

Answer 1

My assumption is that tuples for the same key that contain multiple values should be aggregated. It is unclear what should happen in the same value appears singly and as part of a list, e.g. ("herry", "4"), ("herry, "1,4") however it should be simple enough to remove any such cases
val data = List(("herry","0,1,2"),("herry","1,3"),("herry","3,6"),("herry","4"),("John","5"))// Start writing your ScalaFiddle code here
val (singles, multiples) = data.partition{case (name, list) => !list.contains(",")}
val multiplesAggregated = multiples
.groupBy{case (key, _) => key)
.map{
case (key, values) =>
key -> values.flatMap{case (_, numbers) => numbers.split(",")}.distinct.sorted.mkString(",")}.toList
println(multiplesAggregated ++ singles)
Output:
List((herry,0,1,2,3,6), (herry,4), (John,5))