本文整理汇总了Java中org.apache.cassandra.locator.TokenMetadata.ringIterator方法的典型用法代码示例。如果您正苦于以下问题：Java TokenMetadata.ringIterator方法的具体用法？Java TokenMetadata.ringIterator怎么用？Java TokenMetadata.ringIterator使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类org.apache.cassandra.locator.TokenMetadata的用法示例。
在下文中一共展示了TokenMetadata.ringIterator方法的2个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Java代码示例。
示例1: getRestrictedRanges
import org.apache.cassandra.locator.TokenMetadata; //导入方法依赖的package包/类
/**
* Compute all ranges we're going to query, in sorted order. Nodes can be replica destinations for many ranges,
* so we need to restrict each scan to the specific range we want, or else we'd get duplicate results.
*/
static <T extends RingPosition<T>> List<AbstractBounds<T>> getRestrictedRanges(final AbstractBounds<T> queryRange)
{
// special case for bounds containing exactly 1 (non-minimum) token
if (queryRange instanceof Bounds && queryRange.left.equals(queryRange.right) && !queryRange.left.isMinimum(StorageService.getPartitioner()))
{
return Collections.singletonList(queryRange);
}
TokenMetadata tokenMetadata = StorageService.instance.getTokenMetadata();
List<AbstractBounds<T>> ranges = new ArrayList<AbstractBounds<T>>();
// divide the queryRange into pieces delimited by the ring and minimum tokens
Iterator<Token> ringIter = TokenMetadata.ringIterator(tokenMetadata.sortedTokens(), queryRange.left.getToken(), true);
AbstractBounds<T> remainder = queryRange;
while (ringIter.hasNext())
{
/*
* remainder can be a range/bounds of token _or_ keys and we want to split it with a token:
* - if remainder is tokens, then we'll just split using the provided token.
* - if remainder is keys, we want to split using token.upperBoundKey. For instance, if remainder
* is [DK(10, 'foo'), DK(20, 'bar')], and we have 3 nodes with tokens 0, 15, 30. We want to
* split remainder to A=[DK(10, 'foo'), 15] and B=(15, DK(20, 'bar')]. But since we can't mix
* tokens and keys at the same time in a range, we uses 15.upperBoundKey() to have A include all
* keys having 15 as token and B include none of those (since that is what our node owns).
* asSplitValue() abstracts that choice.
*/
Token upperBoundToken = ringIter.next();
T upperBound = (T)upperBoundToken.upperBound(queryRange.left.getClass());
if (!remainder.left.equals(upperBound) && !remainder.contains(upperBound))
// no more splits
break;
Pair<AbstractBounds<T>,AbstractBounds<T>> splits = remainder.split(upperBound);
if (splits == null)
continue;
ranges.add(splits.left);
remainder = splits.right;
}
ranges.add(remainder);
return ranges;
}
示例2: getRestrictedRanges
import org.apache.cassandra.locator.TokenMetadata; //导入方法依赖的package包/类
/**
* Compute all ranges we're going to query, in sorted order. Nodes can be replica destinations for many ranges,
* so we need to restrict each scan to the specific range we want, or else we'd get duplicate results.
*/
static <T extends RingPosition> List<AbstractBounds<T>> getRestrictedRanges(final AbstractBounds<T> queryRange)
{
// special case for bounds containing exactly 1 (non-minimum) token
if (queryRange instanceof Bounds && queryRange.left.equals(queryRange.right) && !queryRange.left.isMinimum(StorageService.getPartitioner()))
{
return Collections.singletonList(queryRange);
}
TokenMetadata tokenMetadata = StorageService.instance.getTokenMetadata();
List<AbstractBounds<T>> ranges = new ArrayList<AbstractBounds<T>>();
// divide the queryRange into pieces delimited by the ring and minimum tokens
Iterator<Token> ringIter = TokenMetadata.ringIterator(tokenMetadata.sortedTokens(), queryRange.left.getToken(), true);
AbstractBounds<T> remainder = queryRange;
while (ringIter.hasNext())
{
/*
* remainder can be a range/bounds of token _or_ keys and we want to split it with a token:
* - if remainder is tokens, then we'll just split using the provided token.
* - if remainder is keys, we want to split using token.upperBoundKey. For instance, if remainder
* is [DK(10, 'foo'), DK(20, 'bar')], and we have 3 nodes with tokens 0, 15, 30. We want to
* split remainder to A=[DK(10, 'foo'), 15] and B=(15, DK(20, 'bar')]. But since we can't mix
* tokens and keys at the same time in a range, we uses 15.upperBoundKey() to have A include all
* keys having 15 as token and B include none of those (since that is what our node owns).
* asSplitValue() abstracts that choice.
*/
Token upperBoundToken = ringIter.next();
T upperBound = (T)upperBoundToken.upperBound(queryRange.left.getClass());
if (!remainder.left.equals(upperBound) && !remainder.contains(upperBound))
// no more splits
break;
Pair<AbstractBounds<T>,AbstractBounds<T>> splits = remainder.split(upperBound);
if (splits == null)
continue;
ranges.add(splits.left);
remainder = splits.right;
}
ranges.add(remainder);
return ranges;
}