使用 Java 流根据日期间隔对列表数据进行分组并对金额字段求和

Question

Consider there is a class: Partner

Class Partner {
 LocalDate invoiceDate;  
 BigDecimal amount;  
}

Now, I have a list of Partner objects sorted in descending order eg:

[Partner(invoiceDate=2020-01-21, amount=400), 
 Partner(invoiceDate=2020-01-20, amount=400), 
 Partner(invoiceDate=2020-01-11, amount=400), 
 Partner(invoiceDate=2020-01-10, amount=400),
 .....,
 .....]

In the above sample data the field "invoiceDate" is for January month.
note: the list will have data for 12 months or above.

Now,

1. I want to group the data in 15 days interval.ie
   First interval, 1st day to 15th day of the month [1 - 15].
   Second interval, 16th day to Last day of the month [16 - 30/31/28/29].
2. And finally sum the amount value between the date range.

Calculation form the above sample data:
first interval [1-15]: 2 rows qualifies => [invoiceDate=2020-01-11 and invoiceDate=2020-01-10]
second interval [16-31]: 2 rows qualifies => [invoiceDate=2020-01-21 and invoiceDate=2020-01-20]

The final output data should look like this:

[Partner(invoiceDate=2020-01-31, amount=800), 
Partner(invoiceDate=2020-01-15, amount=800)]

note: In the final output invoiceDate should be the last day of the interval.

Answer 1

Using groupingBy would help in such a situation. Here is one of the approaches:

import static java.util.stream.Collectors.*;
Map<String, BigDecimal> dateRangeToSumMap = list
                    .stream()
                    .collect(
                            groupingBy(e -> e.invoiceDate.getDayOfMonth() > 15 ? "16-31" : "1-16",
                                       mapping(
                                               Partner::getAmount, 
                                               reducing(BigDecimal.ZERO, BigDecimal::add)
                                               )
                                       )
                            );

// iterate the map to get the output
dateRangeToSumMap.forEach((k, v) -> {
                
                System.out.println("Date Range = " + k + " , Sum = " + v);
                
            });

Output :

Date Range = 1-16 , Sum = 800
Date Range = 16-31 , Sum = 800

The final output data should look like this:

[Partner(invoiceDate=2020-01-31, amount=800), 
Partner(invoiceDate=2020-01-15, amount=800)]

This can be constructed with the map we have.

note: In the final output invoiceDate should be the last day of the interval.

With year and month , we can get the correct last day of the interval.

Answer 2

You can first map invoiceDate into 15-day interval date. Then use Collectors.toMap to map the data into for invoiceDate and sum the amount in merge function by creating new Parter object. Finally get the map values as list.

Map<LocalDate, Partner> result = list.stream()
    .map(p -> new Partner(p.getInvoiceDate().getDayOfMonth() > 15
        ? p.getInvoiceDate().withDayOfMonth(p.getInvoiceDate().lengthOfMonth())
        : p.getInvoiceDate().withDayOfMonth(15),
          p.getAmount()))
    .collect(Collectors.toMap(Partner::getInvoiceDate, e -> e,
        (a, b) -> new Partner(a.getInvoiceDate(), a.getAmount().add(b.getAmount()))));

List<Partner> res = new ArrayList<>(result.values());

Another approach:

You can simplified the code without creating the Partner object the way @Naman suggested

static LocalDate getIntervalDate(LocalDate d) {
    return (d.getDayOfMonth() > 15 ? d.withDayOfMonth(d.lengthOfMonth()) 
                                   : d.withDayOfMonth(15));
}

List<Partner> res = list.stream()
    .collect(Collectors.toMap(e -> getIntervalDate(e.getInvoiceDate()), 
                              e -> e.getAmount(), (a, b) -> a.add(b)))
    .entrySet()
    .stream()
    .map(e -> new Partner(e.getKey(), e.getValue()))
    .collect(Collectors.toList());

Answer 3

I think this should give you at least something to start with:

Function<Partner, LocalDate> getInterval = (partner) -> {
  // compute if day is before 15th of month and return either 15th
  // of month or last day of month
  return null; 
};

Collection<Partner> partners = Arrays.asList();
Map<LocalDate, BigDecimal> result = partners.stream()
    .collect(Collectors.toMap(getInterval, Partner::getAmount, (a1, a2) -> a1.add(a2)));

You can use the map collector with the merge function to get the sum for all elements with the same key.

Answer 4

Java's verbosity and lack of a tuple type makes the code somewhat longer, this can be done in one pass without any grouping.

public class Main {
    public static void main(String[] args) {
        List<Partner> partners = Arrays.asList(
                new Partner(LocalDate.of(2020, 1, 21), BigDecimal.valueOf(400)),
                new Partner(LocalDate.of(2020, 1, 20), BigDecimal.valueOf(400)),
                new Partner(LocalDate.of(2020, 1, 11), BigDecimal.valueOf(400)),
                new Partner(LocalDate.of(2020, 1, 10), BigDecimal.valueOf(400))
        );

        Accumulator results = IntStream
                .range(0, partners.size())
                .mapToObj(i -> new AbstractMap.SimpleImmutableEntry<>(i, partners.get(i)))
                .reduce(new Accumulator(), (Accumulator acc, Map.Entry<Integer, Partner> e) -> {
                    Partner p = e.getValue();
                    if (p.invoiceDate.getMonthValue() == acc.month || e.getKey() == 0) {
                        BiWeeklyReport bucket = p.invoiceDate.getDayOfMonth() <= 15 ? acc.first : acc.second;
                        bucket.amount = bucket.amount.add(p.amount);
                        bucket.invoiceDate = bucket.invoiceDate.isAfter(p.invoiceDate) ? bucket.invoiceDate : p.invoiceDate;
                        acc.month = bucket.invoiceDate.getMonthValue();
                    }
                    if (e.getKey() == partners.size() - 1 || p.invoiceDate.getMonthValue() != acc.month) {
                        if (acc.first.amount.compareTo(BigDecimal.ZERO) > 0) {
                            acc.reports.add(acc.first);
                        }
                        if (acc.second.amount.compareTo(BigDecimal.ZERO) > 0) {
                            acc.reports.add(acc.second);
                        }
                        acc.first = new BiWeeklyReport();
                        acc.second = new BiWeeklyReport();
                    }
                    return acc;
                }, (acc1, acc2) -> {
                    acc1.reports.addAll(acc2.reports);
                    return acc1;
                });
        System.out.println(results.reports);
    }

    private static class Partner {
        LocalDate invoiceDate;
        BigDecimal amount;

        private Partner(LocalDate invoiceDate, BigDecimal amount) {
            this.invoiceDate = invoiceDate;
            this.amount = amount;
        }
    }

    private static class BiWeeklyReport {
        BigDecimal amount = BigDecimal.ZERO;
        LocalDate invoiceDate = LocalDate.of(1970, 1, 1);

        @Override
        public String toString() {
            return "BiWeeklyReport{" +
                    "amount=" + amount.longValue() +
                    ", invoiceDate=" + DateTimeFormatter.ISO_LOCAL_DATE.format(invoiceDate) +
                    '}';
        }
    }

    private static class Accumulator {
        List<BiWeeklyReport> reports = new ArrayList<>();
        BiWeeklyReport first = new BiWeeklyReport();
        BiWeeklyReport second = new BiWeeklyReport();
        int month = -1;
    }
}

Output:

[BiWeeklyReport{amount=800, invoiceDate=2020-01-11}, BiWeeklyReport{amount=800, invoiceDate=2020-01-21}]