Add option to merge exponential histograms without upscaling

libs/exponential-histogram/src/main/java/org/elasticsearch/exponentialhistogram/ExponentialHistogramMerger.java

@@ -154,6 +154,25 @@ public ExponentialHistogram get() {
         return (result == null) ? ExponentialHistogram.empty() : result;
     }
 
+    /**
+     * Merges the given histogram into the current result, not upscaling it.
+     * This should be used when merging intermediate results to prevent accumulating errors.
+     *
+     * @param toAdd the histogram to merge
+     */
+    public void addWithoutUpscaling(ExponentialHistogram toAdd) {
+        add(toAdd, false);
+    }
+
+    /**
+     * Merges the given histogram into the current result. The histogram might be upscaled if needed.
+     *
+     * @param toAdd the histogram to merge
+     */
+    public void add(ExponentialHistogram toAdd) {
+        add(toAdd, true);
+    }
+
     // This algorithm is very efficient if B has roughly as many buckets as A.
     // However, if B is much smaller we still have to iterate over all buckets of A.
     // This can be optimized by buffering the buckets of small histograms and only merging them when we have enough buckets.
@@ -165,12 +184,7 @@ public ExponentialHistogram get() {
     // It would be possible to only enable the buffering for small histograms,
     // but the optimization seems not worth the added complexity at this point.
 
-    /**
-     * Merges the given histogram into the current result.
-     *
-     * @param toAdd the histogram to merge
-     */
-    public void add(ExponentialHistogram toAdd) {
+    private void add(ExponentialHistogram toAdd, boolean allowUpscaling) {
         ExponentialHistogram a = result == null ? ExponentialHistogram.empty() : result;
         ExponentialHistogram b = toAdd;
 
@@ -193,6 +207,10 @@ public void add(ExponentialHistogram toAdd) {
         // This might involve increasing the scale for B, which would increase its indices.
         // We need to ensure that we do not exceed MAX_INDEX / MIN_INDEX in this case.
         int targetScale = Math.min(maxScale, a.scale());
+        if (allowUpscaling == false) {
+            targetScale = Math.min(targetScale, b.scale());
+        }
+
         if (targetScale > b.scale()) {
             if (negBucketsB.hasNext()) {
                 long smallestIndex = negBucketsB.peekIndex();

libs/exponential-histogram/src/test/java/org/elasticsearch/exponentialhistogram/ExponentialHistogramMergerTests.java

@@ -105,6 +105,14 @@ public void testEmptyZeroBucketIgnored() {
         assertThat(posBuckets.hasNext(), equalTo(false));
     }
 
+    public void testMergeWithoutUpscaling() {
+        ExponentialHistogram histo = createAutoReleasedHistogram(b -> b.scale(0).setPositiveBucket(2, 42));
+        try (ExponentialHistogramMerger merger = ExponentialHistogramMerger.create(100, breaker())) {
+            merger.addWithoutUpscaling(histo);
+            assertThat(merger.get(), equalTo(histo));
+        }
+    }
+
     public void testAggregatesCorrectness() {
         double[] firstValues = randomDoubles(100).map(val -> val * 2 - 1).toArray();
         double[] secondValues = randomDoubles(50).map(val -> val * 2 - 1).toArray();