File indexing completed on 2025-08-28 08:26:53
0001
0002
0003
0004
0005
0006
0007
0008
0009
0010
0011
0012
0013
0014
0015
0016
0017
0018 #pragma once
0019
0020 #include <functional>
0021 #include <memory>
0022 #include <vector>
0023
0024 #include "arrow/acero/accumulation_queue.h"
0025 #include "arrow/acero/bloom_filter.h"
0026 #include "arrow/acero/options.h"
0027 #include "arrow/acero/query_context.h"
0028 #include "arrow/acero/schema_util.h"
0029 #include "arrow/acero/task_util.h"
0030 #include "arrow/result.h"
0031 #include "arrow/status.h"
0032 #include "arrow/type.h"
0033 #include "arrow/util/tracing.h"
0034
0035 namespace arrow {
0036 namespace acero {
0037
0038 using util::AccumulationQueue;
0039
0040 class HashJoinImpl {
0041 public:
0042 using OutputBatchCallback = std::function<Status(int64_t, ExecBatch)>;
0043 using BuildFinishedCallback = std::function<Status(size_t)>;
0044 using FinishedCallback = std::function<Status(int64_t)>;
0045 using RegisterTaskGroupCallback = std::function<int(
0046 std::function<Status(size_t, int64_t)>, std::function<Status(size_t)>)>;
0047 using StartTaskGroupCallback = std::function<Status(int, int64_t)>;
0048 using AbortContinuationImpl = std::function<void()>;
0049
0050 virtual ~HashJoinImpl() = default;
0051 virtual Status Init(QueryContext* ctx, JoinType join_type, size_t num_threads,
0052 const HashJoinProjectionMaps* proj_map_left,
0053 const HashJoinProjectionMaps* proj_map_right,
0054 std::vector<JoinKeyCmp> key_cmp, Expression filter,
0055 RegisterTaskGroupCallback register_task_group_callback,
0056 StartTaskGroupCallback start_task_group_callback,
0057 OutputBatchCallback output_batch_callback,
0058 FinishedCallback finished_callback) = 0;
0059
0060 virtual Status BuildHashTable(size_t thread_index, AccumulationQueue batches,
0061 BuildFinishedCallback on_finished) = 0;
0062 virtual Status ProbeSingleBatch(size_t thread_index, ExecBatch batch) = 0;
0063 virtual Status ProbingFinished(size_t thread_index) = 0;
0064 virtual void Abort(TaskScheduler::AbortContinuationImpl pos_abort_callback) = 0;
0065 virtual std::string ToString() const = 0;
0066
0067 static Result<std::unique_ptr<HashJoinImpl>> MakeBasic();
0068 static Result<std::unique_ptr<HashJoinImpl>> MakeSwiss();
0069
0070 protected:
0071 arrow::util::tracing::Span span_;
0072 };
0073
0074 }
0075 }