1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
|
@Component
@Slf4j
public class BatchProcessor {
private final DataSplitter dataSplitter;
private final ChunkProcessor chunkProcessor;
private final BatchValidator batchValidator;
private final AtomicLong totalBatchesProcessed = new AtomicLong(0);
private final AtomicLong totalItemsProcessed = new AtomicLong(0);
private final AtomicLong totalProcessingTime = new AtomicLong(0);
public <T> List<List<T>> splitData(List<T> data, int batchSize) {
try {
if (data == null || data.isEmpty()) {
return new ArrayList<>();
}
if (batchSize <= 0) {
batchSize = 100;
}
List<List<T>> batches = dataSplitter.split(data, batchSize);
log.debug("数据分割完成: 总数量={}, 批次大小={}, 批次数量={}",
data.size(), batchSize, batches.size());
return batches;
} catch (Exception e) {
log.error("数据分割失败", e);
throw new BatchProcessorException("数据分割失败: " + e.getMessage());
}
}
public <T, R> BatchResult<R> processBatch(List<T> batch, BatchItemProcessor<T, R> processor) {
try {
long startTime = System.currentTimeMillis();
if (!batchValidator.validate(batch)) {
throw new BatchProcessorException("批次验证失败");
}
List<R> results = new ArrayList<>();
List<Exception> exceptions = new ArrayList<>();
int successCount = 0;
int failureCount = 0;
for (int i = 0; i < batch.size(); i++) {
T item = batch.get(i);
try {
R result = processor.process(item);
results.add(result);
successCount++;
} catch (Exception e) {
log.error("处理单个项目失败: 批次索引={}, 项目索引={}", batch.hashCode(), i, e);
exceptions.add(e);
failureCount++;
}
}
long endTime = System.currentTimeMillis();
long processingTime = endTime - startTime;
totalBatchesProcessed.incrementAndGet();
totalItemsProcessed.addAndGet(batch.size());
totalProcessingTime.addAndGet(processingTime);
BatchResult<R> result = BatchResult.<R>builder()
.batchIndex(batch.hashCode())
.successCount(successCount)
.failureCount(failureCount)
.results(results)
.exceptions(exceptions)
.processingTime(processingTime)
.build();
log.debug("批次处理完成: 批次大小={}, 成功数量={}, 失败数量={}, 处理时间={}ms",
batch.size(), successCount, failureCount, processingTime);
return result;
} catch (Exception e) {
log.error("批次处理失败", e);
throw new BatchProcessorException("批次处理失败: " + e.getMessage(), e);
}
}
public <T, R> List<BatchResult<R>> processBatchesParallel(List<List<T>> batches,
BatchItemProcessor<T, R> processor, int parallelism) {
try {
if (batches == null || batches.isEmpty()) {
return new ArrayList<>();
}
if (parallelism <= 0) {
parallelism = Runtime.getRuntime().availableProcessors();
}
ExecutorService executor = Executors.newFixedThreadPool(parallelism);
try {
List<CompletableFuture<BatchResult<R>>> futures = batches.stream()
.map(batch -> CompletableFuture.supplyAsync(() ->
processBatch(batch, processor), executor))
.collect(Collectors.toList());
List<BatchResult<R>> results = futures.stream()
.map(CompletableFuture::join)
.collect(Collectors.toList());
log.info("并行批次处理完成: 批次数量={}, 并行度={}", batches.size(), parallelism);
return results;
} finally {
executor.shutdown();
}
} catch (Exception e) {
log.error("并行批次处理失败", e);
throw new BatchProcessorException("并行批次处理失败: " + e.getMessage(), e);
}
}
public <T, R> List<BatchResult<R>> processBatchesPipeline(List<List<T>> batches,
BatchItemProcessor<T, R> processor, int pipelineSize) {
try {
if (batches == null || batches.isEmpty()) {
return new ArrayList<>();
}
if (pipelineSize <= 0) {
pipelineSize = 3;
}
List<BatchResult<R>> results = new ArrayList<>();
BlockingQueue<List<T>> inputQueue = new LinkedBlockingQueue<>(pipelineSize);
BlockingQueue<BatchResult<R>> outputQueue = new LinkedBlockingQueue<>(pipelineSize);
Thread producer = new Thread(() -> {
try {
for (List<T> batch : batches) {
inputQueue.put(batch);
}
inputQueue.put(new ArrayList<>());
} catch (InterruptedException e) {
Thread.currentThread().interrupt();
}
});
Thread consumer = new Thread(() -> {
try {
List<T> batch;
while ((batch = inputQueue.take()) != null && !batch.isEmpty()) {
BatchResult<R> result = processBatch(batch, processor);
outputQueue.put(result);
}
} catch (InterruptedException e) {
Thread.currentThread().interrupt();
}
});
producer.start();
consumer.start();
BatchResult<R> result;
while ((result = outputQueue.take()) != null) {
results.add(result);
}
producer.join();
consumer.join();
log.info("流水线批次处理完成: 批次数量={}, 流水线大小={}", batches.size(), pipelineSize);
return results;
} catch (Exception e) {
log.error("流水线批次处理失败", e);
throw new BatchProcessorException("流水线批次处理失败: " + e.getMessage(), e);
}
}
public BatchProcessingStatistics getStatistics() {
try {
long totalBatches = totalBatchesProcessed.get();
long totalItems = totalItemsProcessed.get();
long totalTime = totalProcessingTime.get();
return BatchProcessingStatistics.builder()
.totalBatchesProcessed(totalBatches)
.totalItemsProcessed(totalItems)
.totalProcessingTime(totalTime)
.averageBatchSize(totalBatches > 0 ? (double) totalItems / totalBatches : 0)
.averageProcessingTime(totalBatches > 0 ? (double) totalTime / totalBatches : 0)
.build();
} catch (Exception e) {
log.error("获取处理统计信息失败", e);
throw new BatchProcessorException("获取处理统计信息失败: " + e.getMessage());
}
}
}
|
微信
