如何在Java8 Streams中通过多个过滤谓词比较两个Map列表以识别匹配和不匹配的记录

Question

Requirement is to get all the matching and non matching records from the List of Map using multiple matching criteria using the streams. ie Instead of having a single filter to compare only "Email", required to compare two list for matching records with multiple filter predicate for comparing Email and Id both.

List1:

[{"Email","naveen@domain.com", "Id": "A1"}, 
 {"Email":"test@domain.com","id":"A2"}]

List2:

[{"Email","naveen@domain.com", "Id": "A1"}, 
 {"Email":"test@domain.com","id":"A2"}, 
 {"Email":"test1@domain.com","id":"B1"}]

Using streams I'm able to find the matching and non matching records using Single filter predicate on Email: Matching Records :

[{"Email","naveen@domain.com", "Id": "A1"}, 
 {"Email":"test@domain.com","id":"A2"}]

Non Matching Records :

[{"Email":"test1@domain.com","id":"B1"}]]

Is there a way to compare both Email and Id comparison instead of just Email

dbRecords.parallelStream().filter(searchData ->
                inputRecords.parallelStream().anyMatch(inputMap ->
                    searchData.get("Email").equals(inputMap.get("Email")))).
                collect(Collectors.toList());

---

import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.stream.Collectors;
public class ListFiltersToGetMatchingRecords {


    public static void main(String[] args) {

        long startTime = System.currentTimeMillis();
        List<Map<String, Object>> dbRecords = createDbRecords();
        List<Map<String, Object>> inputRecords = createInputRecords();

        List<Map<String,Object>> matchinRecords = dbRecords.parallelStream().filter(searchData ->
                inputRecords.parallelStream().anyMatch(inputMap ->
                    searchData.get("Email").equals(inputMap.get("Email")))).
                collect(Collectors.toList());

        List<Map<String,Object>> notMatchinRecords = inputRecords.parallelStream().filter(searchData ->
                dbRecords.parallelStream().noneMatch( inputMap ->
                        searchData.get("Email").equals(inputMap.get("Email"))
                )).collect(Collectors.toList());

        long endTime = System.currentTimeMillis();
        System.out.println("Matching Records: " + matchinRecords.size());
        matchinRecords.forEach(record -> {
            System.out.println(record.get("Email"));
        });

        System.out.println("Non Matching Records" + notMatchinRecords.size());
        notMatchinRecords.forEach(record -> {
            System.out.println(record.get("Email"));
        });
        System.out.println("Non Matching Records" + notMatchinRecords.size());
        System.out.println("Matching Records: " + matchinRecords.size());
        System.out.println("TotalTImeTaken =" + ((endTime-startTime) /1000) + "sec");
    }

    private static List<Map<String, Object>> createDbRecords() {
        List<Map<String, Object>> dbRecords = new ArrayList<>();
        for(int i =0; i< 100; i+=2) {
            Map<String, Object> dbRecord = new HashMap<>();
            dbRecord.put("Email","naveen" + i +"@gmail.com");
            dbRecord.put("Id", "ID" + i);
            dbRecords.add(dbRecord);
        }
        return dbRecords;
    }

    private static List<Map<String, Object>> createInputRecords() {
        List<Map<String, Object>> dbRecords = new ArrayList<>();
        for(int i =0; i< 100; i++) {
            Map<String, Object> dbRecord = new HashMap<>();
            dbRecord.put("Email", "naveen" + i +"@gmail.com");
            dbRecord.put("ID", "ID" + i);
            dbRecords.add(dbRecord);
        }
        return dbRecords;
    }
}

Answer 1

If you care for performance, you should not combine a linear search with another linear search; with the resulting time complexity can't be fixed with parallel processing when the lists get large.

You should built a data structure which allows efficient lookups first:

Map<List<?>,Map<String, Object>> inputKeys = inputRecords.stream()
    .collect(Collectors.toMap(
        m -> Arrays.asList(m.get("ID"),m.get("Email")),
        m -> m,
        (a,b) -> { throw new IllegalStateException("duplicate "+a+" and "+b); },
        LinkedHashMap::new));

List<Map<String,Object>> matchinRecords = dbRecords.stream()
    .filter(m -> inputKeys.containsKey(Arrays.asList(m.get("ID"),m.get("Email"))))
    .collect(Collectors.toList());

matchinRecords.forEach(m -> inputKeys.remove(Arrays.asList(m.get("ID"),m.get("Email"))));
List<Map<String,Object>> notMatchinRecords = new ArrayList<>(inputKeys.values());

This solution will keep the identity of the Map s.

If you are only interested in the values associated with the "Email" key, it would be much simpler:

Map<Object,Object> notMatchinRecords = inputRecords.stream()
    .collect(Collectors.toMap(
        m -> m.get("ID"),
        m -> m.get("Email"),
        (a,b) -> { throw new IllegalStateException("duplicate"); },
        LinkedHashMap::new));

Object notPresent = new Object();
Map<Object,Object> matchinRecords = dbRecords.stream()
    .filter(m -> notMatchinRecords.getOrDefault(m.get("ID"), notPresent)
                                  .equals(m.get("Email")))
    .collect(Collectors.toMap(
        m -> m.get("ID"),
        m -> m.get("Email"),
        (a,b) -> { throw new IllegalStateException("duplicate"); },
        LinkedHashMap::new));

notMatchinRecords.keySet().removeAll(matchinRecords.keySet());

System.out.println("Matching Records: " + matchinRecords.size());
matchinRecords.forEach((id,email) -> System.out.println(email));

System.out.println("Non Matching Records" + notMatchinRecords.size());
notMatchinRecords.forEach((id,email) -> System.out.println(email));

---

The first variant can get extended to support more/other map entries easily:

List<String> keys = Arrays.asList("ID", "Email");

Function<Map<String,Object>,List<?>> getKey
    = m -> keys.stream().map(m::get).collect(Collectors.toList());

Map<List<?>,Map<String, Object>> inputKeys = inputRecords.stream()
    .collect(Collectors.toMap(
        getKey,
        m -> m,
        (a,b) -> { throw new IllegalStateException("duplicate "+a+" and "+b); },
        LinkedHashMap::new));

List<Map<String,Object>> matchinRecords = dbRecords.stream()
    .filter(m -> inputKeys.containsKey(getKey.apply(m)))
    .collect(Collectors.toList());

matchinRecords.forEach(m -> inputKeys.remove(getKey.apply(m)));
List<Map<String,Object>> notMatchinRecords = new ArrayList<>(inputKeys.values());

Answer 2

You just need to add a condition in the comparison

dbRecords.parallelStream().filter(searchData -> 
                  inputRecords.parallelStream().anyMatch(inputMap ->
                                     searchData.get("Email").equals(inputMap.get("Email"))
                                     && searchData.get("id").equals(inputMap.get("id"))))
         .collect(Collectors.toList());

---

• Add the same in the noneMatch() .
• Compute the average time using System.nanoTime() , it's more accurate
• Try with and without .parallelStream() (just .stream() ) because not sure it helps you)

Answer 3

Here it is mate...

The most efficient way to compare two List of Map to identify the matching and non matching records with multiple filter predicates in Java8 Streams is:

List<Map<String,String>> unMatchedRecords = dbRecords.parallelStream().filter(searchData ->
                inputRecords.parallelStream().noneMatch( inputMap ->
                        searchData.entrySet().stream().noneMatch(value ->
                                inputMap.entrySet().stream().noneMatch(value1 ->
                                        (value1.getKey().equals(value.getKey()) &&
                                                value1.getValue().equals(value.getValue()))))
                )).collect(Collectors.toList());

Note:

1. If <Map<String,String> used above is <Map<Object,Object> instead, don't forget to apply .toString() for .getKey() and value.getKey().

2. The unmatched records thus obtained, could be easily subtracted from either of the list (ie, dbRecords or inputRecords) to retrieve the match results and the operation is swift.

Cheers,

Shubham Chauhan

Answer 4

Why not use && inside anyMatch :

anyMatch(inputMap -> searchData.get("Email").equals(inputMap.get("Email")) 
                     && searchData.get("Id").equals(inputMap.get("Id")))

And I doubt you actually need parallelStream , you do need System.nanoTime on the other hand instead of currentTimeMillis