clohfink commented on code in PR #4589:
URL: https://github.com/apache/cassandra/pull/4589#discussion_r2766638173
##########
src/java/org/apache/cassandra/cql3/restrictions/PartitionKeyRestrictions.java:
##########
@@ -191,6 +193,17 @@ public AbstractBounds<PartitionPosition>
bounds(IPartitioner partitioner, QueryO
*/
private List<ByteBuffer> nonTokenRestrictionValues(QueryOptions options,
ClientState state)
{
+ // fast path, a typical case when a single full restriction is used
+ // for example, when we specify a single partition key to modify
+ if (restrictions.size() == 1 && !restrictions.hasIN())
+ {
+ SingleRestriction r = restrictions.lastRestriction();
+ List<ClusteringElements> values = r.values(options);
+ if (values.size() == 1)
+ return toByteBuffers(Clustering.make(values.get(0).toArray(new
ByteBuffer[comparator.size()])));
Review Comment:
Minor additional possible improvement? For the single-column partition key
case this is allocating:
- 1 × ByteBuffer[] array (unnecessary)
- 1 × BufferClustering object (unnecessary)
to ultimately call get(0) and return the original ByteBuffer. Maybe bypass
Clustering entirely and use ClusteringElements which has it
``` if (values.size() == 1)
{
ClusteringElements elements = values.get(0);
ByteBuffer pk = serializePartitionKey(elements);
validatePartitionKey(pk);
return Collections.singletonList(pk);
}
return toByteBuffers(MultiCBuilder.build(comparator, values));
...
private ByteBuffer serializePartitionKey(ClusteringElements elements)
{
// Single-column partition key: just return the value directly
if (elements.size() == 1)
return elements.get(0);
// Composite partition key: need to build composite
return CompositeType.build(ByteBufferAccessor.instance,
elements.toArray(new
ByteBuffer[elements.size()]));
}
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
