FANNG1 commented on code in PR #10138:
URL: https://github.com/apache/gravitino/pull/10138#discussion_r2875958368
##########
maintenance/optimizer/src/main/java/org/apache/gravitino/maintenance/optimizer/updater/Updater.java:
##########
@@ -102,140 +94,216 @@ public void update(
List<NameIdentifier> nameIdentifiers,
UpdateType updateType) {
StatisticsCalculator calculator =
getStatisticsCalculator(statisticsCalculatorName);
- List<TableMetricWriteRequest> tableMetricWriteRequests = new ArrayList<>();
- List<JobMetricWriteRequest> jobMetricWriteRequests = new ArrayList<>();
+
+ if (UpdateType.STATISTICS.equals(updateType)) {
+ updateStatisticsForIdentifiers(statisticsCalculatorName,
nameIdentifiers, calculator);
+ return;
+ }
+
+ updateMetricsForIdentifiers(statisticsCalculatorName, nameIdentifiers,
calculator);
+ }
+
+ /**
+ * Updates statistics or metrics for all identifiers returned by the
calculator.
+ *
+ * @param statisticsCalculatorName The provider name of the statistics
calculator.
+ * @param updateType The target update type: statistics or metrics.
+ */
+ public void updateAll(String statisticsCalculatorName, UpdateType
updateType) {
+ StatisticsCalculator calculator =
getStatisticsCalculator(statisticsCalculatorName);
+
+ if (UpdateType.STATISTICS.equals(updateType)) {
+ updateAllStatistics(statisticsCalculatorName, calculator);
+ return;
+ }
+
+ updateAllMetrics(statisticsCalculatorName, calculator);
+ }
+
+ @VisibleForTesting
+ public MetricsUpdater getMetricsUpdater() {
+ return metricsUpdater;
+ }
+
+ @Override
+ public void close() throws Exception {
+ closeableGroup.close();
+ }
+
+ private void updateStatisticsForIdentifiers(
+ String statisticsCalculatorName,
+ List<NameIdentifier> nameIdentifiers,
+ StatisticsCalculator calculator) {
+ long tableRecords = 0;
+ long partitionRecords = 0;
+
+ for (NameIdentifier nameIdentifier : nameIdentifiers) {
+ if (!(calculator instanceof SupportsCalculateTableStatistics)) {
+ continue;
+ }
+ SupportsCalculateTableStatistics supportTableStatistics =
+ (SupportsCalculateTableStatistics) calculator;
+ TableAndPartitionStatistics bundle =
+ supportTableStatistics.calculateTableStatistics(nameIdentifier);
+ List<StatisticEntry<?>> statistics = bundle != null ?
bundle.tableStatistics() : List.of();
+ Map<PartitionPath, List<StatisticEntry<?>>> partitionStatistics =
+ bundle != null ? bundle.partitionStatistics() : Map.of();
+
+ tableRecords += countStatistics(statistics);
+ partitionRecords += countPartitionStatistics(partitionStatistics);
+ LOG.info(
+ "Updating table statistics: calculator={}, identifier={}",
+ statisticsCalculatorName,
+ nameIdentifier);
+
+ updateTableStatistics(statistics, nameIdentifier);
+ updatePartitionStatistics(partitionStatistics, nameIdentifier);
+ }
+
+ System.out.println(
+ String.format(
+ "SUMMARY: %s totalRecords=%d tableRecords=%d partitionRecords=%d
jobRecords=%d",
+ UpdateType.STATISTICS.name().toLowerCase(Locale.ROOT),
+ tableRecords + partitionRecords,
+ tableRecords,
+ partitionRecords,
+ 0L));
+ }
+
+ private void updateMetricsForIdentifiers(
+ String statisticsCalculatorName,
+ List<NameIdentifier> nameIdentifiers,
+ StatisticsCalculator calculator) {
+ boolean hasTableMetricsCalculator = calculator instanceof
SupportsCalculateTableMetrics;
+ boolean hasJobMetricsCalculator = calculator instanceof
SupportsCalculateJobMetrics;
+ if (!hasTableMetricsCalculator && !hasJobMetricsCalculator) {
+ throw new IllegalArgumentException(
+ String.format(
+ "Statistics calculator '%s' does not implement metric
interfaces. "
+ + "Expected SupportsCalculateTableMetrics and/or
SupportsCalculateJobMetrics.",
+ statisticsCalculatorName));
+ }
+
long tableRecords = 0;
long partitionRecords = 0;
long jobRecords = 0;
+
for (NameIdentifier nameIdentifier : nameIdentifiers) {
- if (calculator instanceof SupportsCalculateTableStatistics) {
- SupportsCalculateTableStatistics supportTableStatistics =
- ((SupportsCalculateTableStatistics) calculator);
- TableAndPartitionStatistics bundle =
- supportTableStatistics.calculateTableStatistics(nameIdentifier);
- List<StatisticEntry<?>> statistics = bundle != null ?
bundle.tableStatistics() : List.of();
- Map<PartitionPath, List<StatisticEntry<?>>> partitionStatistics =
- bundle != null ? bundle.partitionStatistics() : Map.of();
- tableRecords += countStatistics(statistics);
- partitionRecords += countPartitionStatistics(partitionStatistics);
+ if (hasTableMetricsCalculator) {
+ List<MetricPoint> metrics =
+ ((SupportsCalculateTableMetrics)
calculator).calculateTableMetrics(nameIdentifier);
+ tableRecords += countMetricsByScope(metrics, MetricPoint.Scope.TABLE);
+ partitionRecords += countMetricsByScope(metrics,
MetricPoint.Scope.PARTITION);
LOG.info(
- "Updating table statistics/metrics: calculator={}, updateType={},
identifier={}",
+ "Updating table/partition metrics: calculator={}, identifier={},
count={}",
statisticsCalculatorName,
- updateType,
- nameIdentifier);
- if (UpdateType.STATISTICS.equals(updateType)) {
- updateTableStatistics(statistics, nameIdentifier);
- updatePartitionStatistics(partitionStatistics, nameIdentifier);
- } else {
- tableMetricWriteRequests.addAll(collectTableMetrics(statistics,
nameIdentifier));
- tableMetricWriteRequests.addAll(
- collectPartitionMetrics(partitionStatistics, nameIdentifier));
- }
+ nameIdentifier,
+ metrics == null ? 0 : metrics.size());
+ updateMetrics(metrics);
}
- if (calculator instanceof SupportsCalculateJobStatistics
- && UpdateType.METRICS.equals(updateType)) {
- SupportsCalculateJobStatistics supportJobStatistics =
- ((SupportsCalculateJobStatistics) calculator);
- List<StatisticEntry<?>> statistics =
- supportJobStatistics.calculateJobStatistics(nameIdentifier);
- jobRecords += countStatistics(statistics);
+
+ if (hasJobMetricsCalculator) {
+ List<MetricPoint> metrics =
+ ((SupportsCalculateJobMetrics)
calculator).calculateJobMetrics(nameIdentifier);
+ jobRecords += countMetricsByScope(metrics, MetricPoint.Scope.JOB);
LOG.info(
- "Updating job metrics: calculator={}, identifier={}",
+ "Updating job metrics: calculator={}, identifier={}, count={}",
statisticsCalculatorName,
- nameIdentifier);
- jobMetricWriteRequests.addAll(collectJobMetrics(statistics,
nameIdentifier));
+ nameIdentifier,
+ metrics == null ? 0 : metrics.size());
+ updateMetrics(metrics);
}
Review Comment:
updated
--
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]