In this entry we will explain in depth one of the main collections of Scala the List, today we will indicate some of the most basic aspects that we can do with this type of collection.
The best way to create variables is defining their types, as indicated below, although a less organized way would be without detailing them. In the next code segment for example, line 1 and 4 will have the same value but line 1 is much clearer:
1 2 3 4 | scala> val numbers:List[Int] = List(2, 3, 4, 5, 8, 100, 200, 300) numbers: List[Int] = List(2, 3, 4, 5, 8, 100, 200, 300) scala> val numbers = List(2, 3, 4, 5, 8, 100, 200, 300) numbers: List[Int] = List(2, 3, 4, 5, 8, 100, 200, 300) |
The following lines of code show the creation / console result of lists of String, Int and Lists [Lists] or what are the same lists that in turn contain other lists:
1 2 3 4 5 6 7 8 | scala> val cities:List[String] = List("Havana","Madrid","Oporto") cities: List[String] = List(Havana, Madrid, Oporto) scala> val numbers:List[Int] = List(1,2,4,5) numbers: List[Int] = List(1, 2, 4, 5) scala> val listOfList:List[List[Int]]=List(List(1,2,3),List(0,1,3),List(4,2,1)) listOfList: List[List[Int]] = List(List(1, 2, 3), List(0, 1, 3), List(4, 2, 1)) |
Ways for make a query about if a List is empty:
1 2 3 4 5 6 7 8 | scala> val emptyList = List() emptyList: List[Nothing] = List() scala> emptyList == Nil res3: Boolean = true scala> emptyList.isEmpty res0: Boolean = true |
Since many of the collection methods return a standard Scala type for optional values, we will explain this type in a simple way and in other entries of this blog we already will make a more in-depth explanation of the same one. We speak of the Option type, it is a value to model optional values.
When a value is of type Option it indicates that it can be of two types:
- Some (x) where x is the current value.
- None which indicates that it has no value or has a null value.
In previous sections of code we have seen the simplest way to generate a list. As well as how to check if a list is empty.
A Scala list is composed of 2 parts:
Those who come from other programming languages, the find method of a collection simply passes the object that we want to check if it exists in the collection, the find method of Scala, is different:
The find method receives as parameter a boolean expression that must satisfy an element of the collection so that it can be returned in a Some (x) where x is the found value. In case none item of the list satisfies the boolean expression, it will return a None.
The condition established in line 3 of the previous section indicates that for each element of the cities collection it will be asked if it fulfills the established condition, the first element that fulfills this condition will be returned in a Some (the_looking_element), in case that no element satisfies the condition, as is the case of line 5, will return a None.
The filter method of a collection, as its name implies, given a boolean expression will return all those elements of the collection that meet that condition. An important detail of this method is that if it finds elements that satisfy this condition, those element will be returned in a collection of the same type that we perform the filtering, in case no element meets the condition will return a collection of the type queried empty.
Let's see the following example where there is a list that satisfies the filtered condition and another one in which it does not.
As the following definition indicates, we have the partition method of a collection. For a boolean condition applied to a collection, all those members of the collection that meet the same are grouped on the left side of the tuple and those that do not fulfill on the right part. In short, we will have a tuple contained by two lists.
Let's see an example about a list of integers that we want to divide in 2 groups. The elements > 20 and the rest.
Let's look at our result on line 4 of the previous code segment. Now suppose we want to store those results for process them. Obviously it must be through a tuple, in each part of it we will return a list.
In the left element of the tuple will be those elements that fulfill the boolean condition, in our case the number > 20 and in the right those that do not. Once again the elements will be returned in a collection of the same type that the collections involved in the operation.
A Scala list is composed of 2 parts:
- Block of Elements
- Nil
1 2 3 4 5 | scala> val newList :List[Int]= Nil nuevaLista: List[Int] = List() scala> 1::2::3::4::newList res0: List[Int] = List(1, 2, 3, 4) |
Those who come from other programming languages, the find method of a collection simply passes the object that we want to check if it exists in the collection, the find method of Scala, is different:
def find(p: A => Boolean): Option[A]
The find method receives as parameter a boolean expression that must satisfy an element of the collection so that it can be returned in a Some (x) where x is the found value. In case none item of the list satisfies the boolean expression, it will return a None.
1 2 3 4 5 6 | scala> val cities:List[String] = List("Havana","Madrid","Oporto") cities: List[String] = List(Havana, Madrid, Oporto) scala> cities.find(city=>city.contentEquals("Havana")) res9: Option[String] = Some(Havana) scala> cities.find(city=>city.contentEquals("Havan1a")) res10: Option[String] = None |
The condition established in line 3 of the previous section indicates that for each element of the cities collection it will be asked if it fulfills the established condition, the first element that fulfills this condition will be returned in a Some (the_looking_element), in case that no element satisfies the condition, as is the case of line 5, will return a None.
The filter method of a collection, as its name implies, given a boolean expression will return all those elements of the collection that meet that condition. An important detail of this method is that if it finds elements that satisfy this condition, those element will be returned in a collection of the same type that we perform the filtering, in case no element meets the condition will return a collection of the type queried empty.
def filter(p: (A) ⇒ Boolean): List[A]
Let's see the following example where there is a list that satisfies the filtered condition and another one in which it does not.
1 2 3 4 5 6 | scala> val cities:List[String] = List("Havana","Madrid","Oporto","Malaga","Matanza") cities: List[String] = List(Havana, Madrid, Oporto, Malaga, Matanza) scala> cities.filter(city=>city.startsWith("M")) res2: List[String] = List(Madrid, Malaga, Matanza) scala> cities.filter(city=>city.startsWith("x")) res3: List[String] = List() |
As the following definition indicates, we have the partition method of a collection. For a boolean condition applied to a collection, all those members of the collection that meet the same are grouped on the left side of the tuple and those that do not fulfill on the right part. In short, we will have a tuple contained by two lists.
def partition(p: (A) ⇒ Boolean): (List[A], List[A])
Let's see an example about a list of integers that we want to divide in 2 groups. The elements > 20 and the rest.
1 2 3 4 | scala> val numberList:List[Int] = List(1,3,4,5,8,20,28,14,12) numberList: List[Int] = List(1, 3, 4, 5, 8, 20, 28, 14, 12) scala> numberList.partition(number=>number>10) res3: (List[Int], List[Int]) = (List(20, 28, 14, 12),List(1, 3, 4, 5, 8)) |
Let's look at our result on line 4 of the previous code segment. Now suppose we want to store those results for process them. Obviously it must be through a tuple, in each part of it we will return a list.
In the left element of the tuple will be those elements that fulfill the boolean condition, in our case the number > 20 and in the right those that do not. Once again the elements will be returned in a collection of the same type that the collections involved in the operation.
1 2 3 | scala> val (elementGreaterThan10,elementLessThan10) = numberList.partition(number=>number>10) elementGreaterThan10: List[Int] = List(20, 28, 14, 12) elementLessThan10: List[Int] = List(1, 3, 4, 5, 8) |
In the previous section we leave the result in a tuple whose members (elementGreaterThan10 and elementLessThan10) we can use later in any part of our code.
final def span(p: (A) ⇒ Boolean): (List[A], List[A])
The Span method of a collection, returns a prefix / suffix tuple, from the first element of the collection that does not meet the condition will conform the second part of the tuple, no matter that there are other remaining elements that meet the condition. The simplest way to explain it will be through an example.
scala> val listOfIntegers:List[Int] = List(15, 10, 5, 20, 12,8) listOfIntegers: List[Int] = List(15, 10, 5, 20, 12, 8) scala> listOfIntegers.span(_ < 20) res12: (List[Int], List[Int]) = (List(15, 10, 5),List(20, 12, 8))
As we can see in the previous code segment, we have made a span with those elements less than 20 and has divided our list starting from the number 20(is not less than 20), the first element that did not comply with the condition. No matter that there are other elements after the number 20 that meet the condition, from the first one that fails will form the other part of the tuple. Once again the elements will be returned in a collection of the same type that the collection involved in the operation.
Next we will see the groupBy method, to apply on a collection and that has the following signature:
def groupBy[K](f: (A) ⇒ K): Map[K, List[A]]
This method Partitions the list in a map of lists according to a particular function. So for example if we group a list of String by its length all those elements of the list that have the same length will be in a same key K and will belong to the same List [A].
1 2 3 4 5 | scala> val seqOfString :Seq[String] = Seq("test", "food", "trick", "deal", "canada", "drawing") seqOfString: Seq[String] = List(test, food, trick, deal, canada, drawing) scala> seqOfString.groupBy(_ length) res2: scala.collection.immutable.Map[Int,Seq[String]] = Map(5 -> List(trick), 4 -> List(test, food, deal), 7 -> List(drawing), 6 -> List(canada)) |
In the previous example we have a sequence of String (line 1) and in line 4 we group the items of the collection by the length of the same and we will have as result a Map [Int, List [String]] where each key K is a Length and the value associated with the key is a list that groups all elements of the same length (K). As an important detail, the key K is the type that returns our function used to discriminate, in our previous example we used the length of the string to discriminate so the key is an Integer.
In our next example imagine a list of integers in which we want to group the numbers greater than 20 in one list and the rest in another. So our result will be a Map [Boolean, List [Int]] where we will have 2 keys (K) false and true (line 5). For the key K = true we will have as value a list that groups all the items in the collection greater than 20 and for K = false those less than 20 or equal.
1 2 3 4 5 | scala> val integerList: List[Int] = List(1,2,34,33,56,56,32,20) integerList: List[Int] = List(1, 2, 34, 33, 56, 56, 32, 20) scala> integerList.groupBy(_ > 20) res4: scala.collection.immutable.Map[Boolean,List[Int]] = Map(false -> List(1, 2, 20), true -> List(34, 33, 56, 56, 32)) |
Next we will treat probably one of the most important and most used methods of collections:
def map[B](f: (A) ⇒ B): List[B]
Create a new collection by applying an specific function to each of the elements of the collection on which we want to make the map. Once again the elements will be returned in a collection of the same type that the collection involved in the operation.
1 2 3 4 5 6 7 8 | scala> val listForMap:List[Int] =List(10,20,30,40,50) listForMap: List[Int] = List(10, 20, 30, 40, 50) scala> val multipliedList = listForMap.map(_ * 2) multipliedList: List[Int] = List(20, 40, 60, 80, 100) scala> val multipliedList = listForMap.map(element => element * 2) multipliedList: List[Int] = List(20, 40, 60, 80, 100) |
The previous code segment shows how a new list is created by applying a function to each of the items in the collection. In line 4 multiplies each element of the list x 2, the result of this line and the one of line 7 is the same only that in line 4 we use what we call placeholder, later we will see as depending on one utility or another It will be more convenient to use one option or another.
Let's take a more direct look at the use of functions. We will put the simplest possible example of applying functions, but it can be applied to any function. Remember that we always return a collection of the same type as the collection on which we perform the operation.
In the next code segment line 1 and line 4 are shown in two different ways to define a function that multiplies the input parameter x 2. Remember that listForMap was already defined in the previous code segment and now we will see how are applied the different functions to each elements of the collection.
1 2 3 4 5 6 7 8 9 10 11 | scala> def flong(x:Int):Int= x*2 flong: (x: Int)Int scala> def f: Int => Int = x => x*2 f: Int => Int scala> val multipliedList = listForMap.map(element => f(element)) multipliedList: List[Int] = List(20, 40, 60, 80, 100) scala> val multipliedList = listForMap.map(element => flong(element)) multipliedList: List[Int] = List(20, 40, 60, 80, 100) |
In next entries I will put examples of a greater complexity in the use of this method although the explained here can be extrapolated to any situation.
We will see another important method in a collection. We refer to flatMap. This method constructs a new collection by applying a function to all items in the list and using the elements of the resulting collection. In general it makes sense in nested collections, we talk about collections that contain other collections and we want our resulting collection to be the most external collection, although we may also occasionally seek the opposite or that the resulting collection is the innermost. Either way is better to see it through examples.
def flatMap[B](f: (A) ⇒ GenTraversableOnce[B]): List[B]
The following example defines a list that contains a list of cities or a list in which each one of its items is in turn another list of String. As the following example indicates:
1 2 3 4 5 | scala> val ciudades: List[List[String]]=List(List("madrid","barcelona"),List("havana","cienfuegos"),List("london","manchester")) ciudades: List[List[String]] = List(List(madrid, barcelona), List(havana, cienfuegos), List(london, manchester)) scala> ciudades.flatMap(x=>x.map(cities=>cities.toUpperCase)) res1: List[String] = List(MADRID, BARCELONA, HAVANA, CIENFUEGOS, LONDON, MANCHESTER) |
We apply a function to each of the elements (cities) of the list and the result is left in a new collection. Notice how the internal collections have been processed and their elements have been taken to the container collection, it is always the same with the flatMaps, if instead of each Items being a list of String we were in front of certain object, we could act on each one of the objects and return in the container collection the elements resulting from the application of a certain function on the items in the collection.
Let's see another example to see it from another point of view:
1 2 3 4 5 | scala> val listaDeString: List[String] = List("madrid","londres","paris") listaDeStrings: List[String] = List(madrid, london, paris) scala> listaDeStrings.flatMap(char=>char.toUpperCase) res9: List[Char] = List(M, A, D, R, I, D, L, O, N, D, O,N, P, A, R, I, S) |
We process the elements and result pull them out to a container collection. Here each of the elements will be each character that make up the array of string that form our words. Remember that we always return a collection of the same type that we perform the operation.
Now that we have seen both the flatMap and the map, let's see an example that combines both, where a List of Maps will return a List of a Map, all have been joined in the container collection (in other entries, the Map will be explained as Collection, we speak of a key-value structure in a quick way):
1 2 3 4 5 6 7 8 9 10 11 12 13 14 | scala> val seqMapNumbers: Seq[Map[String,Int]] = List(Map("Six" -> 6, "five" -> 5, "four" -> 4),Map("Nine" -> 9, "eight" -> 8, "seven" -> 7)) seqMapNumbers: Seq[Map[String,Int]] = List(Map(Six -> 6, five -> 5, four -> 4), Map(Nine -> 9, eight -> 8, seven -> 7)) scala> seqMapNumbers.flatMap((x=>x)) res3: Seq[(String, Int)] = List((Six,6), (five,5), (four,4), (Nine,9), (eight,8), (seven,7)) scala> seqMapNumbers.flatMap((x=>x)).map(y=>y._1) res4: Seq[String] = List(Six, five, four, Nine, eight, seven) scala> seqMapNumbers.flatMap((mapa=>mapa)).map(mapa=>mapa._1) res5: Seq[String] = List(Six, five, four, Nine, eight, seven) scala> seqMapNumbers.flatMap((mapa=>mapa)).map(mapa=>mapa._2) res6: Seq[Int] = List(6, 5, 4, 9, 8, 7) |
According to the previous code segment we have as input a Seq [Map [String, Int]] , a sequence where each Item is a Map and when applied flatMap on it we have a list of Tuples or what is the same a List of One Map. All items that were in individual maps have been removed and can now be processed as a common Map.
Let's look at the following code section:
1 2 3 4 5 | scala> seqMapNumbers.flatMap(x=>x) res9: Seq[(String, Int)] = List((Six,6), (five,5), (four,4), (Nine,9), (eight,8), (seven,7)) scala> seqMapNumbers.flatMap(x=>x).toMap res8: scala.collection.immutable.Map[String,Int] = Map(four -> 4, Nine -> 9, seven -> 7, five -> 5, eight -> 8, Six -> 6) |
With regard to what we explained above, we see how in line 4 applying a map to our resulting flatMap we obtain a Map of all the elements. We started with Seq [Map [String, Int]] and got Map [String, Int], in this case we have opened the content to our inner collection.
We have headOption, method that returns the first element of the collection, was used in combination with filter, so that we filter a collection and if the filtered collection we apply the headOption it returns the first element of the collection in an Some(first_element) or a None as no element meets the requirements.
def headOption: Option[A]
Imagine in the next code segment a list of integers (line 1). Now we filter the list for those elements that fulfill a certain condition and then we get the first element of the subset already filtered (line 4). As a detail, if the list is empty (line 7) and instead of a headOption we had a head method would give us an exception, this is the great advantage of our headOption.
scala> val listOfNumber: List[Int] = List(10,30,45,76,66,20) listOfNumber: List[Int] = List(10, 30, 45, 76, 66, 20) scala> val optionOfHeadOfNumber: Option[Int] = listOfNumber.filter(_ > 30).headOption optionOfHeadOfNumber: Option[Int] = Some(45) scala> val optionOfHeadOfNumber: Option[Int] = listOfNumber.filter(_ > 100).headOption optionOfHeadOfNumber: Option[Int] = None
A solution similar to the previous but more complete comes with the find method.
def find(p: (A) ⇒ Boolean): Option[A]
Find the first element, if it exists and satisfies the Boolean condition. Our previous combination (filter + headOption) is done in a single method, achieving the same result and the consequent advantage of its use, in a line do almost everything.
scala> val optionOfHeadOfNumber: Option[Int] = listOfNumber.find(_ > 30) optionOfHeadOfNumber: Option[Int] = Some(45)
Let's see the operation of the fold method which "merge or synthesize" the elements of the collection using an associative operator (which can be executed in an arbitrary order [sum, multiplication]). Always have an initial value or accumulator.
def fold[A1 >: A](z: A1)(op: (A1, A1) ⇒ A1): A1
It is important to emphasize that in the fold the operations on the elements set of the collection do not have an order and besides the accumulator (z) can be applied as parameter to the operator (op) an indeterminate amount of times, as indicated by the specification .
Let's look the following example:
scala> List(1,2,3).fold("gdg")( _.toString() + _.toString())
1st iteration -> "gdg" + "1" -> "gdg1"
2nd iteration -> "gdg1" + "2" -> "gdg12"
3rd iteration -> "gdg12" + "3" -> "gdg123"
["possible fourth iteration"] -> "gdg123" + "gdg" -> "gdg123gdg"
And so on. In all tests that I have performed the behavior has been similar until the "3rd iteration" but if we consider the Scala language specification the result can be as indicated until the fourth iteration and even could be continue exiting iterations. In addition, the order of operations may NOT be as indicated above. That is why our accumulator is recommended to be a neutral value (0 for integer values, empty string for String values, etc).
Let's see in the following example how to "merge" or synthesize a list starting from "zero" so that our final result is the sum of all the elements of the list. Our accumulator is the basis of the function that we will apply to the collection with which as a final result we will get the sum of the elements of the list + the sum of our accumulator one or n times, just as recommended by the specification, if this element is neutral (it is the case) will not affect the final result.
1 2 3 4 5 6 7 8 9 10 11 12 13 | scala> val listOfInteger: List[Int] = List (1,2,3,4,5) listOfInteger: List[Int] = List(1, 2, 3, 4, 5) scala> (listOfInteger fold 0) (_ * _) res11: Int = 0 scala> (listOfInteger fold 0) (_ + _) res12: Int = 15 scala> listOfInteger.fold (0) (_ + _) res13: Int = 15 scala> listOfInteger.fold (10) (_ + _) |
Lines 7 and 10 are two ways of expressing the same thing, which in our case is reduced to adding the elements of an integer list.
We have several details to highlight:
Always the result of the previous sum is added to the next element and our accumulator in the lines that interest us is 0, can be passed as a parameter of the operation at any point in the process. However the result of an operation like the present one in line 13 would be 25 (10 [our accumulator] + 15). In both cases (line 10 and 13) we do not obtain a collection, as the signature of the method indicates, we obtain the list "merged" or "synthesized" as a result of an associative operation (in this case sum).
Let's see the following example to leave us in a clearer way the functionality of this method.
1 2 3 4 5 | scala> ((listOfInteger) fold "") ((s1, s2) => s"$s1 - $s2") res15: Any = " - 1 - 2 - 3 - 4 - 5" scala> ((listOfInteger) fold "test") ((s1, s2) => s"$s1 - $s2") res16: Any = test - 1 - 2 - 3 - 4 - 5 |
As we can see in the previous code segment, now our accumulator is a string. See the difference when the accumulator is an empty string and when it has the value "test". We have pending the different types of data in Scala so in this moment not to worry about the type Any that appears lines 2 and 5, although they are of great importance in this case, considering the signature of the method. The fact that the accumulator (test) is not a neutral value can result in other random values different from what is shown in the previous section of code.
The following operation will generate an error because it is an operation in which if decisive the order of the operations to be performed:
scala> List("1","2","3").fold(0)(_ + _.toInt) <console>:11: error: value toInt is not a member of Any List("1","2","3").fold(0)(_ + _.toInt)
It is important to note that in situations like this (where the order of operations is important) we can not consider the use of fold. Based on the previous line of code would be an error the following analysis:
1st iteration -> op (0, "1") -> 0 + "1" .toInt -> 1
2nd iteration -> op (1, "2") -> 1 + "2" .toInt -> 3
3rd iteration -> op (3, "3") -> 3 + "3" .toInt -> 6
Starting from the fact that the fold nature itself indicates that there is no set order for operations then NOTHING indicates that the first iteration is op (0, "1") -> 0 + "1" .toInt -> 1 could just be any other combination in which for example is op (0, "1") -> 1 + 0.toInt -> 1 and we would already have our first error.
We must know an important detail of the functionality of fold depending the object that apply this method will not be the same the concept of "synthesize or merge" in all types of collections or objects on which fold method can be applied, so we will assume here that the concept of polymorphism is known by all.
Some of the above errors can be solved with our next method.
def foldLeft[B](z: B)(op: (B, A) ⇒ B): B
As noted in the specification op (... op (z, x_1), x_2, ..., x_n) we will apply the operator in a direction from LEFt -> RIGht
So that from LEFt to RIGht:
1st iteration op (accumulator, first_element_element) -> Result_1
2nd iteration op (Result_1, second_element_element_collection) -> Result_2
3rd iteration op (Result_2, third_element_element_collection) -> Result_3
And so on until the collection ends.
One of the big differences between foldLeft and fold is that in fold there is no guarantee on the order in which the elements are processed while in foldLeft we know that we start from LEFt -> RIGht.
Let's look at a case study where we use foldLeft and however fold method can not be used:
scala> List("1","2","3").foldLeft(0)(_ + _.toInt) res59: Int = 6
The sequence order of the operations executed here is as follows:
1st iteration (0 + "1" .toInt) -> 1
2nd iteration (1 + "2" .toInt) -> 3
3rd iteration (3 + "3" .toInt) -> 6 [the last result]
We will always have the guarantee that any foldLeft we make on a collection the first operation applied on the collection will be op (accumulator, 1st_collection_item).
Our next method will be the foldRight whose operation, speaking in the simplest way will be similar to foldLeft but both the iterations and the order of the initial accumulator will be contrary to foldLeft, we will start from right to left..
def foldRight[B](z: B)(op: (A, B) ⇒ B): B
As we can appreciate the signature of the method is the same although the way to do the iterations is RIGht -> LEFt
We have a very simple example but that will give us an idea:
scala> List(1,2,3,4,5,6,7,8,9,10).foldRight(1)((a,b) => {println(s"$a,$b");a}) 10,1 9,10 8,9 7,8 6,7 5,6 4,5 3,4 2,3 1,2 res12: Int = 1
Let's look at another similar example(see foldLeft section):
scala> List("1","2","3").foldRight(0)( _.toInt + _) res14: Int = 6
We see how has been necessary to change the order of the parameters. And now let's see the order of the operations.
Important: See the order or position that the accumulator (on the right) occupies in our case.
1st iteration op (first_element_element, accumulator) -> Result_1
2nd iteration op (second_element_element, Result_1) -> Result_2
3rd iteration op (third_element_element_, Result_2) -> Result_3
A common operation of all folds is that if the collection is empty we return the accumulator, which in general we must try to be NEUTRAL.
What we mean by NEUTRAL:
If it's a list -> Nil
If it is a multiplication operator -> 1
If it is a String operator -> ""
And so on.
If we can adapt our code to use fold would be perfect although many times we will not be able to. If we can adapt our code instead of using foldRight using foldLeft it would be much better(foldLeft on foldRight). We will explain it in future entries.
We will now deal the reduce method, whose main difference with the fold method is that it does not have an initial value, so we have to make sure that the collection on which we use it is not an empty collection. Also here is "reduced or merged" a collection using associative operators.
def reduce[A1 >: A](op: (A1, A1) ⇒ A1): A1
Let's take some example like when we want to generate composite csv files obviously for several CSV entries. Given the following collection let's see how foldLeft resolves generate a file whose lines have values separated by commas:
scala> val values: List[String] = List("ValueCSV1", "ValueCSV2", "ValueCVS3","ValueCSV4","ValueCSV5") values: List[String] = List(ValueCSV1, ValueCSV2, ValueCVS3, ValueCSV4, ValueCSV5) scala> values.fold("")(_ + "," + _) res11: String = ,ValueCSV1,ValueCSV2,ValueCVS3,ValueCSV4,ValueCSV5
If we process our fold method we can delete the "," but if we do not want to do this processing we could run a reduce method on the collection:
scala> values.reduce(_ + "," + _) res12: String = ValueCSV1,ValueCSV2,ValueCVS3,ValueCSV4,ValueCSV5 scala> values.reduceLeft(_ + "," + _) res13: String = ValueCSV1,ValueCSV2,ValueCVS3,ValueCSV4,ValueCSV5
All you have to guarantee is that values are not an empty list. In those occasions that reduceLeft and reduceRight methods lead us to the same result the most advisable is to use reduce, the most common situation. As we explained previously for fold method, when we need the order of the operations to be a specific order we will have to use the appropriate (left or right) reduce method.
scala> listNumber.reduce(_ min _) res31: Int = 2 scala> listNumber.reduce(_ max _) res32: Int = 20 scala> listNumber.reduceLeft(_ max _) res33: Int = 20 scala> listNumber.reduceRight(_ max _) res34: Int = 20 scala> listNumber.max res35: Int = 20 scala> listNumber.min
This is a case where it would be better to use reduce or neither that. But imagine that our op instead of being max or min, is a division, in this case we must care about the order of the parameters and we would then have to see what reduces method (left or right) to use.
Find, flatten, diff, contains, exists FilterNot, foreach, last, tail, collect and above all arithmetic operations on list (::, :::, +:, ++:, ++ and rest of this kinds of op)
Anyway, before doing any operation better look at the Api, in most of the times there are always different routes, some more optimal than others.
Entradas
