Github user paul-rogers commented on a diff in the pull request:
https://github.com/apache/drill/pull/1107#discussion_r167283942
--- Diff:
exec/java-exec/src/main/java/org/apache/drill/exec/record/AbstractRecordBatchMemoryManager.java
---
@@ -0,0 +1,63 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.drill.exec.record;
+
+import org.apache.drill.exec.physical.impl.spill.RecordBatchSizer;
+import org.apache.drill.exec.vector.ValueVector;
+
+public abstract class AbstractRecordBatchMemoryManager {
+ private int outgoingRowWidth;
+ private int outputRowCount = ValueVector.MAX_ROW_COUNT;
+ protected static final int OFFSET_VECTOR_WIDTH = 4;
+ protected static final int WORST_CASE_FRAGMENTATION_FACTOR = 2;
+ protected static final int MAX_NUM_ROWS = ValueVector.MAX_ROW_COUNT;
+ protected static final int MIN_NUM_ROWS = 1;
+
+ public void update(int inputIndex) {};
+
+ public void update() {};
+
+ public int getOutputRowCount() {
+ return outputRowCount;
+ }
+
+ /**
+ * Given batchSize and rowWidth, this will set output rowCount taking
into account
+ * the min and max that is allowed.
+ */
+ public void setOutputRowCount(long batchSize, int rowWidth) {
+ this.outputRowCount = Math.min(ValueVector.MAX_ROW_COUNT,
+
Math.max(RecordBatchSizer.safeDivide(batchSize/WORST_CASE_FRAGMENTATION_FACTOR,
rowWidth), MIN_NUM_ROWS));
--- End diff --
I suspect the math may be off here. `batchSize` is the actual memory for
the batch? If so, then it already includes actual internal fragmentation which,
depending on the source, can be 25% (for "good" operators) to 50% (if from the
network) to 95% (pathological Parquet cases.) Using the fragmentation factor on
this value gets us even more lost.
A better way is to do:
```
rawBatchSize = rowWidth * rowCount;
actualBatchSize = rowBatchSize * FRAG_FACTOR;
```
Or to compute the output size:
```
targetOutputSize = // some value you set
targetDataSize = targetOutputSize / FRAG_FACTOR
targetRowCount = targetDataSize / rowWidth
```
---
