Accumulator not working properly in parallel stream - java

I made collector who can reduce a stream to a map which has the keys as the items that can be bought by certain customers and the names of customers as values, my implementation is working proberly in sequential stream
but when i try to use parallel it's not working at all, the resulting sets always contain one customer name.
List<Customer> customerList = this.mall.getCustomerList();
Supplier<Object> supplier = ConcurrentHashMap<String,Set<String>>::new;
BiConsumer<Object, Customer> accumulator = ((o, customer) -> customer.getWantToBuy().stream().map(Item::getName).forEach(
item -> ((ConcurrentHashMap<String,Set<String>>)o)
.merge(item,new HashSet<String>(Collections.singleton(customer.getName())),
(s,s2) -> {
HashSet<String> res = new HashSet<>(s);
res.addAll(s2);
return res;
})
));
BinaryOperator<Object> combiner = (o,o2) -> {
ConcurrentHashMap<String,Set<String>> res = new ConcurrentHashMap<>((ConcurrentHashMap<String,Set<String>>)o);
res.putAll((ConcurrentHashMap<String,Set<String>>)o2);
return res;
};
Function<Object, Map<String, Set<String>>> finisher = (o) -> new HashMap<>((ConcurrentHashMap<String,Set<String>>)o);
Collector<Customer, ?, Map<String, Set<String>>> toItemAsKey =
new CollectorImpl<>(supplier, accumulator, combiner, finisher, EnumSet.of(
Collector.Characteristics.CONCURRENT,
Collector.Characteristics.IDENTITY_FINISH));
Map<String, Set<String>> itemMap = customerList.stream().parallel().collect(toItemAsKey);
There is certainly a problem in my accumulator implementation or another Function but I cannot figure it out! could anyone suggest what should i do ?

Your combiner is not correctly implemented.
You overwrite all entries that has the same key. What you want is adding values to existing keys.
BinaryOperator<ConcurrentHashMap<String,Set<String>>> combiner = (o,o2) -> {
ConcurrentHashMap<String,Set<String>> res = new ConcurrentHashMap<>(o);
o2.forEach((key, set) -> set.forEach(string -> res.computeIfAbsent(key, k -> new HashSet<>())
.add(string)));
return res;
};

Related

creating Map from List is not giving expected result

I have the two list objects as shown below, from which i'm creating the map object.
List<Class1> list1;
List<Class2> list2;
HashMap<String,String> map = new HashMap<>();
for(Class1 one : list1){
if(one.isStatus()){
map.put(one.getID(),one.getName());
}
}
//iterating second list
for(Class2 two : list2){
if(two.isPerformed()){
map.put(two.getID(),two.getName());
}
}
The above code works fine , want the above to be written using streams.
Below is the sample code using streams().
map = list1.stream().filter(one.isStatus()).collect(toMap(lst1 -> lst1.getID(), lst1.getName());
map = list2.stream().filter(...);
But the "map" is not giving the expected result when written using stream() API.
Stream concatenation Stream.concat may be applied here to avoid map.putAll
Map<String, String> map = Stream.concat(
list1.stream()
.filter(Class1::isStatus)
.map(obj -> Arrays.asList(obj.getID(), obj.getName())),
list2.stream()
.filter(Class2::isPerformed)
.map(obj -> Arrays.asList(obj.getID(), obj.getName()))
) // Stream<List<String>>
.collect(Collectors.toMap(
arr -> arr.get(0), // key - ID
arr -> arr.get(1),
(v1, v2) -> v1 // keep the first value in case of possible conflicts
));
The code above uses a merge function (v1, v2) -> v1 to handle possible conflicts when the same ID occurs several times in list1 and/or list2 to keep the first occurrence.
However, the following merge function allows joining all the occurrences into one string value (v1, v2) -> String.join(", ", v1, v2).
I'm not sure what expected result you're not seeing but I created a minimal working example that you should be able to adapt for your own use case.
public class Main {
public static void main(String[] args) {
List<Person> personList = new ArrayList<>();
Map<Integer, String> personMap = personList.stream()
.filter(Person::isStatus)
.collect(Collectors.toMap(person -> person.id, person -> person.name));
}
private static class Person {
public String name;
public int id;
public boolean isStatus() {
return true;
}
}
}
Try this,
List<Class1> list1;
List<Class2> list2;
Map<String, String> map1 = list1.stream().filter(Class1::isStatus).collect(Collectors.toMap(Class1::getId, Class1::getName));
Map<String, String> map2 = list2.stream().filter(Class2::isPerformed).collect(Collectors.toMap(Class2::getId, Class2::getName));
map1.putAll(map2);

Can't convert tuple list to hashmap java

I want to convert a javax.persistence.Tuple into a HashMap, but like this, it inserts the last element of the tuple and takes also the alias and data type. How can I improve this method so it takes values of the tuple?
public Map<String, Object> tuplesToMap(List<Tuple> data){
Map<String, Object> values = new HashMap<>();
data.forEach(tuple -> {
tuple.getElements().forEach(
element -> {
values.put(element.getAlias(), tuple.get(element));
}
);
});
return values;
}
with java 8 is simply :
return data.stream()
.collect(Collectors.toMap(
t -> t.get(0, String.class),
t -> t.get(1, Object.class)));
Seems to be working :
public static List<Map<String/*UPPERCASE*/, Object>> jpaTuplesToMaps(
List<javax.persistence.Tuple> data
){
return data.stream()
.map(tuple -> { // per each tuple of the input List
// creating a new HashMap
Map<String, Object> resultItem = new HashMap<>();
// filling the created HashMap with values of
tuple.getElements().forEach( // each column of the tuple
col -> { resultItem.put(col.getAlias(), tuple.get(col)); }
);
// returning the created HashMap instead of the current Tuple
return resultItem;
})
// collecting & returning all the created HashMap-s as a List
.collect(Collectors.toList());
}
But usualy both single & list conversions are required, so let's combine them :
public static Map<String/*UPPERCASE*/, Object> jpaTupleToMap(
javax.persistence.Tuple data /*CASE INSENSITIVE*/
){
Map<String, Object> result =
new HashMap<>(); // exactly HashMap since it can handle NULL keys & values
data.getElements().forEach(
col -> { result.put(col.getAlias(), data.get(col)); }
);
return result;
}
//-------------------------
public static List<Map<String/*UPPERCASE*/, Object>> jpaTuplesToMaps(
List<javax.persistence.Tuple> data /*CASE INSENSITIVE*/
){
return data.stream() // List<Tuple> -> Tuple1,..TupleN
.map(tuple -> jpaTupleToMap(tuple)) // Tuple1 -> HashMap1,..TupleN -> HashMapN
.collect(Collectors.toList()); // HashMap1,..HashMapN -> List
}
The element.getAlias() you're using as the key for the hashmap is probably same for some of the elements.
Map keys are unique, meaning, if you insert entries (1, "one") and then (1, "two"), the first value will be overridden by the latter. If you want to have multiple values mapped to one key, use Map<String, Collection<Object>>, or a Multimap from Guava, which is exactly the same thing.
You can insert into multimap with this function - if the key is not in the map, create a new ArrayList and add it to the map, otherwise return the existing one. Then, insert the value to the list:
values
.computeIfAbsent(element.getAlias, k -> new ArrayList<>())
.add(tuple.get(element));

How to convert Stream<Map<Integer, String>> to map java 8

Here I am posting sample datastructure
I have a list List<Result> resultsList;
class Result {
String name;
Map<String,Integer> resultMap;
}
Now I would like to stream through this list and get the map.
resultList.stream().filter(result->"xxx".equals(result.getName()))
.map(result->result.getResultMap);
It returns Stream<Map<String,Integer>> but I need only Map<String,Integer>.
How to get it using java 8 streams?
Update:
As geneqew mentioned
This is how my datastructure looks
List<Result> resultsList;
Map<String, Integer> map1 = new HashMap<>();
map1.put("m1", 1);
Map<String, Integer> map2 = new HashMap<>();
map2.put("m2", 2);
Map<String, Integer> map3 = new HashMap<>();
map3.put("m3", 3);
results = Arrays.asList(
new Result("r1", map1),
new Result("r2", map2),
new Result("r3", map3)
);
I would like to retrieve single map based on name.
for (Result result: resultsList)
{
if ('xxx'.equals(result.getName())
{
return result.getResultMap();
}
}
Since you want to return the result map of the first Result element to pass your filter, you can obtain it with findFirst():
Optional<Map<String,Integer>> resultMap =
resultList.stream()
.filter(result->"xxx".equals(result.getName()))
.map(Result::getResultMap)
.findFirst();
You can extract the Map from the Optional this way:
Map<String,Integer> resultMap =
resultList.stream()
.filter(result->"xxx".equals(result.getName()))
.map(Result::getResultMap)
.findFirst()
.orElse(null);
if you're only looking for one item:
resultList.stream()
.filter(result -> "xxx".equals(result.getName()))
.map(Result::getResultMap)
.findAny();
if the filter could match more than one item then you'll need to flatten then toMap it:
resultList.stream()
.filter(result-> "xxx".equals(result.getName()))
.flatMap(result -> result.getResultMap().entrySet().stream())
.collect(toMap(Map.Entry::getKey, Map.Entry::getValue));
if there can be duplicates then use the merge function to resolve collisions:
resultList.stream()
.filter(result -> "xxx".equals(result.getName()))
.flatMap(result -> result.getResultMap().entrySet().stream())
.collect(toMap(Map.Entry::getKey, Map.Entry::getValue, (l, r) -> l));
Since you only wanted the map that matches the results' name then:
results.stream()
.filter(r-> r.getName().equals("r2"))
.map(r-> r.getResultMap())
.findFirst()
.orElse(null);
given you have a sample content of:
List<Result> results;
Map<String, Integer> map1 = new HashMap<>();
map1.put("m1", 1);
Map<String, Integer> map2 = new HashMap<>();
map2.put("m2", 2);
Map<String, Integer> map3 = new HashMap<>();
map3.put("m3", 3);
results = Arrays.asList(
new Result("r1", map1),
new Result("r2", map2),
new Result("r3", map3)
);
A bit of explanation, you got a stream because the last operation in your stream is a map; assuming in your list its possible to have more than 1 result with the same name, findFirst will return the first match if found otherwise an empty optional is returned; Finally orElse to get terminate the stream, providing a null value on empty match.
So I want to explain why you receive stream and not a map. The reason of this is because in the beginning you have List with Result objects that you filter by some criteria (in your case "xxx".equals(result.getName())).
Now you can have as result zero, one or more elements that will pass this criteria! Java does not know how many elements will pass at compile time and that is why you get Stream.
Imagine situation that you have two Result objects that have the same name 'xxx' then you will have two maps. The question is what you want to do? If you get only one of the maps you will loose information. If you want to get all of them, please try something like this:
List<Map<String,Integer>> listWithResultMaps = resultList.stream()
.filter(result->"xxx".equals(result.getName()))
.map(result->result.getResultMap())
.collect(Collectors.toList());
Now in this listWithResultMaps you can process all maps that you have as result of your filter.
Good Luck!

Iterate big hashmap in parallel

I have a linked hashmap which may contain upto 300k records at maximum. I want to iterate this map in parallel to improve the performance. The function iterates through the map of vectors and finds dot product of given vector against all the vectors in map. Also have one more check based on date value. And the function returns a nested hashmap. T
This is the code using iterator:
public HashMap<String,HashMap<String,Double>> function1(String key, int days) {
LocalDate date = LocalDate.now().minusDays(days);
HashMap<String,Double> ret = new HashMap<>();
HashMap<String,Double> ret2 = new HashMap<>();
OpenMapRealVector v0 = map.get(key).value;
for(Map.Entry<String, FixedTimeHashMap<OpenMapRealVector>> e: map.entrySet()) {
if(!e.getKey().equals(key)) {
Double d = v0.dotProduct(e.getValue().value);
d = Double.parseDouble(new DecimalFormat("###.##").format(d));
ret.put(e.getKey(),d);
if(e.getValue().date.isAfter(date)){
ret2.put(e.getKey(),d);
}
}
}
HashMap<String,HashMap<String,Double>> result = new HashMap<>();
result.put("dot",ret);
result.put("anomaly",ret2);
return result;
}
Update:
I looked into Java 8 streams, but I am running into CastException and Null pointer exceptions when using the parallel stream as this map is being modified else where.
Code:
public HashMap<String,HashMap<String,Double>> function1(String key, int days) {
LocalDate date = LocalDate.now().minusDays(days);
HashMap<String,Double> ret = new HashMap<>();
HashMap<String,Double> ret2 = new HashMap<>();
OpenMapRealVector v0 = map.get(key).value;
synchronized (map) {
map.entrySet().parallelStream().forEach(e -> {
if(!e.getKey().equals(key)) {
Double d = v0.dotProduct(e.getValue().value);
d = Double.parseDouble(new DecimalFormat("###.##").format(d));
ret.put(e.getKey(),d);
if(e.getValue().date.isAfter(date)) {
ret2.put(e.getKey(),d);
}
}
});
}
}
I have synchronized the map usage, but it still gives me the following errors:
java.util.concurrent.ExecutionException: java.lang.ClassCastException
Caused by: java.lang.ClassCastException
Caused by: java.lang.ClassCastException: java.util.HashMap$Node cannot be cast to java.util.HashMap$TreeNode
Also, I was thinking Should i split up the map into multiple pieces and run each using different threads in parallel?
You need to retrieve the Set<Map.Entry<K, V>> from the map.
Here's how you iterate on a Map using parallel Streams in Java8:
Map<String, String> myMap = new HashMap<> ();
myMap.entrySet ()
.parallelStream ()
.forEach (entry -> {
String key = entry.getKey ();
String value = entry.getValue ();
// here add whatever processing you wanna do using the key / value retrieved
// ret.put (....);
// ret2.put (....)
});
Clarification:
The maps ret and ret2 should be declared as ConcurrentHashMaps to allow the concurrent inserts / updates from multiple threads.
So the declaration of the 2 maps become:
Map<String,Double> ret = new ConcurrentHashMap<> ();
Map<String,Double> ret2 = new ConcurrentHashMap<> ();
One possible solution using Java 8 would be,
Map<String, Double> dotMap = map.entrySet().stream().filter(e -> !e.getKey().equals(key))
.collect(Collectors.toMap(Map.Entry::getKey, e -> Double
.parseDouble(new DecimalFormat("###.##").format(v0.dotProduct(e.getValue().value)))));
Map<String, Double> anomalyMap = map.entrySet().stream().filter(e -> !e.getKey().equals(key))
.filter(e -> e.getValue().date.isAfter(date))
.collect(Collectors.toMap(Map.Entry::getKey, e -> Double
.parseDouble(new DecimalFormat("###.##").format(v0.dotProduct(e.getValue().value)))));
result.put("dot", dotMap);
result.put("anomaly", anomalyMap);
Update
Here's much more elegant solution,
Map<String, Map<String, Double>> resultMap = map.entrySet().stream().filter(e -> !e.getKey().equals(key))
.collect(Collectors.groupingBy(e -> e.getValue().date.isAfter(date) ? "anomaly" : "dot",
Collectors.toMap(Map.Entry::getKey, e -> Double.parseDouble(
new DecimalFormat("###.##").format(v0.dotProduct(e.getValue().value))))));
Here we first group them based on anomaly or dot, and then use a downstream Collector to create a Map for each group. Also I have updated .filter() criteria based on the following suggestions.

Java 8 Join Map with Collectors.toMap

I'm trying to collect in a Map the results from the process a list of objects and that it returns a map. I think that I should do it with a Collectors.toMap but I haven't found the way.
This is the code:
public class Car {
List<VersionCar> versions;
public List<VersionCar> getVersions() {
return versions;
}
}
public class VersionCar {
private String wheelsKey;
private String engineKey;
public String getWheelsKey() {
return wheelsKey;
}
public String getEngineKey() {
return engineKey;
}
}
process method:
private static Map<String,Set<String>> processObjects(VersionCar version) {
Map<String,Set<String>> mapItems = new HashMap<>();
mapItems.put("engine", new HashSet<>(Arrays.asList(version.getEngineKey())));
mapItems.put("wheels", new HashSet<>(Arrays.asList(version.getWheelsKey())));
return mapItems;
}
My final code is:
Map<String,Set<String>> mapAllItems =
car.getVersions().stream()
.map(versionCar -> processObjects(versionCar))
.collect(Collectors.toMap()); // here I don't know like collect the map.
My idea is to process the list of versions and in the end get a Map with two items: wheels and engine but with a set<> with all different items for all versions. Do you have any ideas as can I do that with Collectors.toMap or another option?
The operator you want to use in this case is probably "reduce"
car.getVersions().stream()
.map(versionCar -> processObjects(versionCar))
.reduce((map1, map2) -> {
map2.forEach((key, subset) -> map1.get(key).addAll(subset));
return map1;
})
.orElse(new HashMap<>());
The lambda used in "reduce" is a BinaryOperator, that merges 2 maps and return the merged map.
The "orElse" is just here to return something in the case your initial collection (versions) is empty.
From a type point of view it gets rid of the "Optional"
You can use Collectors.toMap(keyMapper, valueMapper, mergeFunction). Last argument is used to resolve collisions between values associated with the same key.
For example:
Map<String, Set<String>> mapAllItems =
car.getVersions().stream()
.map(versionCar -> processObjects(versionCar))
.flatMap(m -> m.entrySet().stream())
.collect(Collectors.toMap(Map.Entry::getKey, Map.Entry::getValue,
(firstSet, secondSet) -> {
Set<String> result = new HashSet<>();
result.addAll(firstSet);
result.addAll(secondSet);
return result;
}
));
To get the mapAllItems, we don't need and should not define processObjects method:
Map<String, Set<String>> mapAllItems = new HashMap<>();
mapAllItems.put("engine", car.getVersions().stream().map(v -> v.getEngineKey()).collect(Collectors.toSet()));
mapAllItems.put("wheels", car.getVersions().stream().map(v -> v.getWheelsKey()).collect(Collectors.toSet()));
Or by AbstractMap.SimpleEntry which is lighter than the Map created byprocessObjects`:
mapAllItems = car.getVersions().stream()
.flatMap(v -> Stream.of(new SimpleEntry<>("engine", v.getEngineKey()), new SimpleEntry<>("wheels", v.getWheelsKey())))
.collect(Collectors.groupingBy(e -> e.getKey(), Collectors.mapping(e -> e.getValue(), Collectors.toSet())));

Categories

Resources