ClickHouse/dbms/include/DB/Storages/MergeTree/MergeSelector.h

#pragma once

#include <cstddef>
#include <vector>
#include <functional>


namespace DB
{

/** Interface of algorithm to select data parts to merge
 *   (merge is also known as "compaction").
  * Following properties depend on it:
  *
  * 1. Number of data parts at some moment in time.
  *    If parts are merged frequently, then data will be represented by lower number of parts, in average,
  *     but with cost of higher write amplification.
  *
  * 2. Write amplification ratio: how much times, on average, source data was written
  *     (during initial writes and followed merges).
  *
  * Number of parallel merges are controlled outside of scope of this interface.
  */
class IMergeSelector
{
public:
	/// Information about data part relevant to merge selecting strategy.
	struct Part
	{
		/// Size of data part in bytes.
		size_t size;

		/// How old this data part in seconds.
		time_t age;

		/// Depth of tree of merges by which this part was created. New parts has zero level.
		unsigned level;

		/// Opaque pointer to avoid dependencies (it is not possible to do forward declaration of typedef).
		const void * data;
	};

	/// Parts are belong to partitions. Only parts within same partition could be merged.
	using PartsInPartition = std::vector<Part>;

	/// Parts are in some specific order. Parts could be merged only in contiguous ranges.
	using Partitions = std::vector<PartsInPartition>;

	/** Function could be called at any frequency and it must decide, should you do any merge at all.
	  * If better not to do any merge, it returns empty result.
	  */
	virtual PartsInPartition select(
		const Partitions & partitions,
		const size_t max_total_size_to_merge) = 0;

	virtual ~IMergeSelector() {}
};

}