我目前正在尝试将Java 8的Stream API合并到我的日常Java工具箱中。我正在尝试使用Streams来查找正整数的素因子,然后将每个因子存储在数组(或ArrayList
)中,并将它们的多重性存储在并行数组中。或者,我正在尝试创建一个说明... FactorWithMultiplicity
对象的流,或者甚至是Map
,其中因子为关键,多重性为值。如果因子按升序排序,如果它甚至可以处理非常大的数字(例如,我敢说Long.MAX_VALUE
),那就太好了。
目前,我的代码看起来像这样,但是,由于我是Streams的初学者,我确信有更快或更适合的方式来完成此任务。请使用Streams创建您的解决方案,但如果您知道某些非Stream解决方案更快,请随时指向该代码。
int num = getPositiveInt();
ArrayList<Integer> factors = new ArrayList<>();
ArrayList<Integer> multiplicities = new ArrayList<>();
boolean isPrime = IntStream.rangeClosed(2, num / 2)
.reduce(num, (int temp, int factor) -> {
int count = 0;
while (temp % factor == 0) {
temp /= factor;
count++;
}
if (count > 0) {
factors.add(factor);
multiplicities.add(count);
}
return temp;
}) > 1;
答案 0 :(得分:4)
如果您特别需要基于流的解决方案,则可以使用递归方式对数字进行因子分析:
IntStream factors(int num) {
return IntStream.range(2, num)
.filter(x -> num % x == 0)
.mapToObj(x -> IntStream.concat(IntStream.of(x), factors(num / x)))
.findFirst()
.orElse(IntStream.of(num));
}
然后您可以使用以下代码制作两个列表:
Map<Integer, Integer> f2m = factors(2, num).boxed()
.collect(toMap(f -> f, f -> 1, Integer::sum)); // or groupingBy with summingInt(f->1), whichever you prefer
List<Integer> factors = new ArrayList<>(f2m.keySet());
List<Integer> multiplicities = factors.stream().map(f2m::get).collect(toList());
如果您希望从中获得更多性能,可以将最后找到的因素传递给factors
方法并使用该方法而不是2
。
如果您想要考虑多头,那么这里有一些性能改进的版本:
static LongStream factors(long lastFactor, long num) {
return LongStream.rangeClosed(lastFactor, (long) Math.sqrt(num))
.filter(x -> num % x == 0)
.mapToObj(x -> LongStream.concat(LongStream.of(x), factors(x, num / x)))
.findFirst()
.orElse(LongStream.of(num));
}
如果您希望结果按排序顺序排列,可以使用
SortedMap<Long, Integer> f2m = factors(2, num).boxed()
.collect(toMap(f -> f, f -> 1, Integer::sum, TreeMap::new));
或者,保持Map
不变,然后使用
List<Long> factors = f2m.keySet().stream().sorted().collect(toList());
答案 1 :(得分:1)
如果您想反复调用factorsOf
,那么另一种变体很有用。 (我在某个地方偷了筛子的基本想法,修好了。)
这里的想法是使用素数作为流,过滤因子并确定它们的多样性,以创建建立结果的FactorTimes对象。
public class PrimeFactors {
private final int limit = 1_000_000;
private BitSet sieve = new BitSet( limit+1 );
public PrimeFactors(){
sieve.set( 2, limit );
long count = sieve.stream()
.peek( x -> { if( (long)x*x < limit )
for( int i = x*x; i <= limit; i += x )
sieve.clear( i );
})
.count();
}
public FactorTimes[] factorsOf( int num ){
FactorTimes[] fts = sieve.stream()
.limit( num/2 )
.filter( x -> num % x == 0 )
.mapToObj( x -> { int n = 1;
int k = num/x;
while( k % x == 0 ){ k /= x; n++; }
return new FactorTimes( x, n );
} )
.toArray( FactorTimes[]::new );
return fts;
}
public static void main( String[] args ){
PrimeFactors pf = new PrimeFactors();
for( FactorTimes ft: pf.factorsOf( 4504500 ) ){
System.out.println( ft );
}
}
}
class FactorTimes {
private int factor, multiplicity;
public FactorTimes(int f, int m) {
factor = f; multiplicity = m;
}
public int getFactor() { return factor; }
public int getMultiplicity() { return multiplicity; }
public String toString(){
return multiplicity > 1 ? factor + "(" + multiplicity + ")"
: Integer.toString( factor ); }
}
答案 2 :(得分:1)
要生成素数因子,您需要跟踪几种状态。因此Streams
不适合这项任务。
您可以做的是提供自己的Spliterator
来创建IntStream
。现在您可以生成数组或分组操作:
public static IntStream primeFactors(int n) {
int characteristics = Spliterator.ORDERED | Spliterator.SORTED | Spliterator.IMMUTABLE | Spliterator.NONNULL;
Spliterator.OfInt spliterator = new Spliterators.AbstractIntSpliterator(Long.MAX_VALUE, characteristics) {
int val = n;
int div = 2;
@Override
public boolean tryAdvance(IntConsumer action) {
while (div <= val) {
if (val % div == 0) {
action.accept(div);
val /= div;
return true;
}
div += div == 2 ? 1 : 2;
}
return false;
}
@Override
public Comparator<? super Integer> getComparator() {
return null;
}
};
return StreamSupport.intStream(spliterator, false);
}
并打电话给这样的话:
int n = 40500;
System.out.println(Arrays.toString(primeFactors(n).toArray()));
System.out.println(primeFactors(n).boxed().collect(
Collectors.groupingBy(Function.identity(), Collectors.summingInt(i -> 1)))
);
你应该得到你想要的结果:
[2, 2, 3, 3, 3, 3, 5, 5, 5]
{2=2, 3=4, 5=3}
答案 3 :(得分:0)
当分解整数时,最有利可图的优化是尝试除数直到数字的平方根(包括,例如:尝试分解49)。此外,在检查2之后,您可以在事后检查只有奇数。
int num = getPositiveInt();
ArrayList<Integer> factors = new ArrayList<>();
ArrayList<Integer> multiplicities = new ArrayList<>();
int factor = 2;
int f_delta = 1; // to increase by +1 only once (2 to 3)
while ((factor*factor)<=num) {
int count = 0;
while (num % factor == 0) {
num /= factor;
count++;
}
if (count > 0) {
factors.add(factor);
multiplicities.add(count);
}
factor += f_delta;
f_delta = 2;
}
答案 4 :(得分:0)
在彻底调查之后,我发现与我在问题中发布的内容相比,这是速度的极大提升。唯一更快的是 @Misha 在更改factors
功能后使用.rangeClosed(prevFactor, Math.sqrt(num))
发布的更快的内容。但是, this other solution 是最快的解决方案...期间......但不使用Streams。
public static Map<Long, Integer> factorize(long num) { //NOW USING LONG
Map<Long, Integer> factors = new LinkedHashMap<>(); //NOW USING MAP
long lastRemainder = LongStream.rangeClosed(2, (long) Math.sqrt(num)) //NOW USING SQRT
.filter(x -> (x== 2||x%2>0)&&(x==3||x%3>0)&&(x==5||x%5>0)) //ADDED THIS
.reduce(num, (temp, factor) -> {
if (factor <= temp / factor) { //ADDED THIS
int count = 0;
while (temp % factor == 0) {
temp /= factor;
count++;
}
if (count > 0)
factors.put(factor, count);
}
return temp;
});
if (lastRemainder != num && lastRemainder > 1) //ADDED THIS
factors.put(lastRemainder, 1);
return factors;
}