Github user arunmahadevan commented on a diff in the pull request:
https://github.com/apache/storm/pull/1950#discussion_r123982153
--- Diff:
storm-core/src/jvm/org/apache/storm/state/BaseBinaryStateIterator.java ---
@@ -0,0 +1,162 @@
+package org.apache.storm.state;
+
+import com.google.common.collect.Iterators;
+import com.google.common.collect.PeekingIterator;
+import com.google.common.primitives.UnsignedBytes;
+
+import java.util.AbstractMap;
+import java.util.Arrays;
+import java.util.Iterator;
+import java.util.Map;
+import java.util.NoSuchElementException;
+import java.util.Set;
+import java.util.TreeSet;
+
+/**
+ * Base implementation of iterator over {@link KeyValueState} which is
based on binary type.
+ */
+public abstract class BaseBinaryStateIterator<K, V> implements
Iterator<Map.Entry<K, V>> {
+
+ private final PeekingIterator<Map.Entry<byte[], byte[]>>
pendingPrepareIterator;
+ private final PeekingIterator<Map.Entry<byte[], byte[]>>
pendingCommitIterator;
+ private final Set<byte[]> providedKeys;
+
+ private boolean firstLoad = true;
+ private PeekingIterator<Map.Entry<byte[], byte[]>> pendingIterator;
+ private PeekingIterator<Map.Entry<byte[], byte[]>> cachedResultIterator;
+
+ /**
+ * Constructor.
+ *
+ * @param pendingPrepareIterator The iterator of pendingPrepare
+ * @param pendingCommitIterator The iterator of pendingCommit
+ */
+ public BaseBinaryStateIterator(Iterator<Map.Entry<byte[], byte[]>>
pendingPrepareIterator,
+ Iterator<Map.Entry<byte[], byte[]>> pendingCommitIterator) {
+ this.pendingPrepareIterator =
Iterators.peekingIterator(pendingPrepareIterator);
+ this.pendingCommitIterator =
Iterators.peekingIterator(pendingCommitIterator);
+ this.providedKeys = new
TreeSet<>(UnsignedBytes.lexicographicalComparator());
+ }
+
+ @Override
+ public boolean hasNext() {
+ if (seekToAvailableEntry(pendingPrepareIterator)) {
+ pendingIterator = pendingPrepareIterator;
+ return true;
+ }
+
+ if (seekToAvailableEntry(pendingCommitIterator)) {
+ pendingIterator = pendingCommitIterator;
+ return true;
+ }
+
+
+ if (firstLoad) {
+ // load the first part of entries
+ fillCachedResultIterator();
+ firstLoad = false;
+ }
+
+ while (true) {
+ if (seekToAvailableEntry(cachedResultIterator)) {
+ pendingIterator = cachedResultIterator;
+ return true;
+ }
+
+ if (isEndOfDataFromStorage()) {
+ break;
+ }
+
+ fillCachedResultIterator();
+ }
+
+ pendingIterator = null;
+ return false;
+ }
+
+ private void fillCachedResultIterator() {
+ Iterator<Map.Entry<byte[], byte[]>> iterator =
loadChunkFromStateStorage();
+ if (iterator != null) {
+ cachedResultIterator = Iterators.peekingIterator(iterator);
+ } else {
+ cachedResultIterator = null;
+ }
+ }
+
+ @Override
+ public Map.Entry<K, V> next() {
+ if (!hasNext()) {
+ throw new NoSuchElementException();
+ }
+
+ Map.Entry<byte[], byte[]> keyValue = pendingIterator.next();
+
+ K key = decodeKey(keyValue.getKey());
+ V value = decodeValue(keyValue.getValue());
+
+ providedKeys.add(keyValue.getKey());
+ return new AbstractMap.SimpleEntry(key, value);
+ }
+
+ @Override
+ public void remove() {
+ throw new UnsupportedOperationException();
+ }
+
+ /**
+ * Load some part of state KVs from storage and returns iterator of
cached data from storage.
+ *
+ * @return Iterator of loaded state KVs
+ */
+ protected abstract Iterator<Map.Entry<byte[],byte[]>>
loadChunkFromStateStorage();
+
+ /**
+ * Check whether end of data is reached from storage state KVs.
+ *
+ * @return whether end of data is reached from storage state KVs
+ */
+ protected abstract boolean isEndOfDataFromStorage();
+
+ /**
+ * Decode key to convert byte array to state key type.
+ *
+ * @param key byte array encoded key
+ * @return Decoded value of key
+ */
+ protected abstract K decodeKey(byte[] key);
+
+ /**
+ * Decode value to convert byte array to state value type.
+ *
+ * @param value byte array encoded value
+ * @return Decoded value of value
+ */
+ protected abstract V decodeValue(byte[] value);
--- End diff --
Can we let the subclass deal with the specifics of decoding than assuming
data is encoded as `byte[]`?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
