[
https://issues.apache.org/jira/browse/BEAM-3446?focusedWorklogId=145667&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-145667
]
ASF GitHub Bot logged work on BEAM-3446:
----------------------------------------
Author: ASF GitHub Bot
Created on: 19/Sep/18 14:08
Start Date: 19/Sep/18 14:08
Worklog Time Spent: 10m
Work Description: vvarma commented on a change in pull request #5841:
[BEAM-3446] Fixes RedisIO non-prefix read operations
URL: https://github.com/apache/beam/pull/5841#discussion_r218818530
##########
File path:
sdks/java/io/redis/src/main/java/org/apache/beam/sdk/io/redis/RedisIO.java
##########
@@ -280,28 +316,77 @@ public void processElement(ProcessContext
processContext) throws Exception {
while (!finished) {
ScanResult<String> scanResult = jedis.scan(cursor, scanParams);
List<String> keys = scanResult.getResult();
-
- Pipeline pipeline = jedis.pipelined();
- if (keys != null) {
- for (String key : keys) {
- pipeline.get(key);
- }
- List<Object> values = pipeline.syncAndReturnAll();
- for (int i = 0; i < values.size(); i++) {
- processContext.output(KV.of(keys.get(i), (String) values.get(i)));
- }
+ for (String k : keys) {
+ processContext.output(k);
}
-
cursor = scanResult.getStringCursor();
if ("0".equals(cursor)) {
finished = true;
}
}
}
+ }
+ /** A {@link DoFn} requesting Redis server to get key/value pairs. */
+ private static class ReadFn extends BaseReadFn<KV<String, String>> {
+ @Nullable transient Multimap<BoundedWindow, String> bundles = null;
+ @Nullable AtomicInteger batchCount = null;
+ private final int batchSize;
+
+ @StartBundle
+ public void startBundle(StartBundleContext context) {
+ bundles = ArrayListMultimap.create();
+ batchCount = new AtomicInteger();
+ }
- @Teardown
- public void teardown() {
- jedis.close();
+ ReadFn(RedisConnectionConfiguration connectionConfiguration, int
batchSize) {
+ super(connectionConfiguration);
+ this.batchSize = batchSize;
+ }
+
+ private int getBatchSize() {
+ return batchSize;
+ }
+
+ @ProcessElement
+ public void processElement(ProcessContext processContext, BoundedWindow
window)
+ throws Exception {
+ String key = processContext.element();
+ bundles.put(window, key);
+ if (batchCount.incrementAndGet() > getBatchSize()) {
+ Multimap<BoundedWindow, KV<String, String>> kvs = fetchAndFlush();
+ for (BoundedWindow w : kvs.keySet()) {
+ for (KV<String, String> kv : kvs.get(w)) {
+ processContext.output(kv);
+ }
+ }
+ }
+ }
+
+ private Multimap<BoundedWindow, KV<String, String>> fetchAndFlush() {
+ Multimap<BoundedWindow, KV<String, String>> kvs =
ArrayListMultimap.create();
+ for (BoundedWindow w : bundles.keySet()) {
+ String[] keys = new String[bundles.get(w).size()];
+ bundles.get(w).toArray(keys);
+ List<String> results = jedis.mget(keys);
+ for (int i = 0; i < results.size(); i++) {
+ if (results.get(i) != null) {
+ kvs.put(w, KV.of(keys[i], results.get(i)));
+ }
+ }
+ }
+ bundles = ArrayListMultimap.create();
+ batchCount.set(0);
+ return kvs;
+ }
+
+ @FinishBundle
+ public void finishBundle(FinishBundleContext context) throws Exception {
+ Multimap<BoundedWindow, KV<String, String>> kvs = fetchAndFlush();
Review comment:
Thanks @iemejia . Do suggest if there any other changes needed.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
Issue Time Tracking
-------------------
Worklog Id: (was: 145667)
Time Spent: 6h (was: 5h 50m)
> RedisIO non-prefix read operations
> ----------------------------------
>
> Key: BEAM-3446
> URL: https://issues.apache.org/jira/browse/BEAM-3446
> Project: Beam
> Issue Type: New Feature
> Components: io-java-redis
> Reporter: Vinay varma
> Assignee: Vinay varma
> Priority: Major
> Time Spent: 6h
> Remaining Estimate: 0h
>
> Read operation in RedisIO is for prefix based look ups. While this can be
> used for exact key matches as well, the number of operations limits the
> through put of the function.
> I suggest exposing current readAll operation as readbyprefix and using more
> simpler operations for readAll functionality.
> ex:
> {code:java}
> String output = jedis.get(element);
> if (output != null) {
> processContext.output(KV.of(element, output));
> }
> {code}
> instead of:
> https://github.com/apache/beam/blob/7d240c0bb171af6868f1a6e95196c9dcfc9ac640/sdks/java/io/redis/src/main/java/org/apache/beam/sdk/io/redis/RedisIO.java#L292
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
