Develop Reference

How does Flinks Collector.collect() handle data?

Im trying to understand what Flinks Collector.collect() does and how it handles incoming/outgoing data:
Example taken from Flink DataSet API:
The following code transforms a DataSet of text lines into a DataSet of words:
DataSet<String> output = input.flatMap(new Tokenizer());
public class Tokenizer implements FlatMapFunction<String, String> {
#Override
public void flatMap(String value, Collector<String> out) {
for (String token : value.split("\\W")) {
out.collect(token);
}
}
}
So the text Lines get split into tokens and each of them gets "collected". As intuitive as it might sound but im missing the actual dynamics behind Collector.collect(). Where is the collected data stored before it gets assigned to output i.e does Flink put them in some sort of Buffer? And if yes, how is the data transferred to the network?

from the official source code documentation.
Collects a record and forwards it. The collector is the "push"
counterpart of the {#link java.util.Iterator}, which "pulls" data in.
So, it receives a value and stores one or more values into the Iterator. Then pushes to the next operator. But this is a matter of the network stack/ buffers.

Apache Beam in Dataflow Large Side Input

This is most similar to this question.
I am creating a pipeline in Dataflow 2.x that takes streaming input from a Pubsub queue. Every single message that comes in needs to be streamed through a very large dataset that comes from Google BigQuery and have all the relevant values attached to it (based on a key) before being written to a database.
The trouble is that the mapping dataset from BigQuery is very large - any attempt to use it as a side input fails with the Dataflow runners throwing the error "java.lang.IllegalArgumentException: ByteString would be too long". I have attempted the following strategies:
1) Side input
As stated,the mapping data is (apparently) too large to do this. If I'm wrong here or there is a work-around for this, please let me know because this would be the simplest solution.
2) Key-Value pair mapping
In this strategy, I read the BigQuery data and Pubsub message data in the first part of the pipeline, then run each through ParDo transformations that change every value in the PCollections to KeyValue pairs. Then, I run a Merge.Flatten transform and a GroupByKey transform to attach the relevant mapping data to each message.
The trouble here is that streaming data requires windowing to be merged with other data, so I have to apply windowing to the large, bounded BigQuery data as well. It also requires that the windowing strategies are the same on both datasets. But no windowing strategy for the bounded data makes sense, and the few windowing attempts I've made simply send all the BQ data in a single window and then never send it again. It needs to be joined with every incoming pubsub message.
3) Calling BQ directly in a ParDo (DoFn)
This seemed like a good idea - have each worker declare a static instance of the map data. If it's not there, then call BigQuery directly to get it. Unfortunately this throws internal errors from BigQuery every time (as in the entire message just says "Internal error"). Filing a support ticket with Google resulted in them telling me that, essentially, "you can't do that".
It seems this task doesn't really fit the "embarrassingly parallelizable" model, so am I barking up the wrong tree here?
EDIT :
Even when using a high memory machine in dataflow and attempting to make the side input into a map view, I get the error java.lang.IllegalArgumentException: ByteString would be too long
Here is an example (psuedo) of the code I'm using:
Pipeline pipeline = Pipeline.create(options);
PCollectionView<Map<String, TableRow>> mapData = pipeline
.apply("ReadMapData", BigQueryIO.read().fromQuery("SELECT whatever FROM ...").usingStandardSql())
.apply("BQToKeyValPairs", ParDo.of(new BQToKeyValueDoFn()))
.apply(View.asMap());
PCollection<PubsubMessage> messages = pipeline.apply(PubsubIO.readMessages()
.fromSubscription(String.format("projects/%1$s/subscriptions/%2$s", projectId, pubsubSubscription)));
messages.apply(ParDo.of(new DoFn<PubsubMessage, TableRow>() {
#ProcessElement
public void processElement(ProcessContext c) {
JSONObject data = new JSONObject(new String(c.element().getPayload()));
String key = getKeyFromData(data);
TableRow sideInputData = c.sideInput(mapData).get(key);
if (sideInputData != null) {
LOG.info("holyWowItWOrked");
c.output(new TableRow());
} else {
LOG.info("noSideInputDataHere");
}
}
}).withSideInputs(mapData));
The pipeline throws the exception and fails before logging anything from within the ParDo.
Stack trace:
java.lang.IllegalArgumentException: ByteString would be too long: 644959474+1551393497
com.google.cloud.dataflow.worker.repackaged.com.google.protobuf.ByteString.concat(ByteString.java:524)
com.google.cloud.dataflow.worker.repackaged.com.google.protobuf.ByteString.balancedConcat(ByteString.java:576)
com.google.cloud.dataflow.worker.repackaged.com.google.protobuf.ByteString.balancedConcat(ByteString.java:575)
com.google.cloud.dataflow.worker.repackaged.com.google.protobuf.ByteString.balancedConcat(ByteString.java:575)
com.google.cloud.dataflow.worker.repackaged.com.google.protobuf.ByteString.balancedConcat(ByteString.java:575)
com.google.cloud.dataflow.worker.repackaged.com.google.protobuf.ByteString.copyFrom(ByteString.java:559)
com.google.cloud.dataflow.worker.repackaged.com.google.protobuf.ByteString$Output.toByteString(ByteString.java:1006)
com.google.cloud.dataflow.worker.WindmillStateInternals$WindmillBag.persistDirectly(WindmillStateInternals.java:575)
com.google.cloud.dataflow.worker.WindmillStateInternals$SimpleWindmillState.persist(WindmillStateInternals.java:320)
com.google.cloud.dataflow.worker.WindmillStateInternals$WindmillCombiningState.persist(WindmillStateInternals.java:951)
com.google.cloud.dataflow.worker.WindmillStateInternals.persist(WindmillStateInternals.java:216)
com.google.cloud.dataflow.worker.StreamingModeExecutionContext$StepContext.flushState(StreamingModeExecutionContext.java:513)
com.google.cloud.dataflow.worker.StreamingModeExecutionContext.flushState(StreamingModeExecutionContext.java:363)
com.google.cloud.dataflow.worker.StreamingDataflowWorker.process(StreamingDataflowWorker.java:1000)
com.google.cloud.dataflow.worker.StreamingDataflowWorker.access$800(StreamingDataflowWorker.java:133)
com.google.cloud.dataflow.worker.StreamingDataflowWorker$7.run(StreamingDataflowWorker.java:771)
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
java.lang.Thread.run(Thread.java:745)

Check out the section called "Pattern: Streaming mode large lookup tables" in Guide to common Cloud Dataflow use-case patterns, Part 2. It might be the only viable solution since your side input doesn't fit into memory.
Description:
A large (in GBs) lookup table must be accurate, and changes often or
does not fit in memory.
Example:
You have point of sale information from a retailer and need to
associate the name of the product item with the data record which
contains the productID. There are hundreds of thousands of items
stored in an external database that can change constantly. Also, all
elements must be processed using the correct value.
Solution:
Use the "Calling external services for data enrichment" pattern
but rather than calling a micro service, call a read-optimized NoSQL
database (such as Cloud Datastore or Cloud Bigtable) directly.
For each value to be looked up, create a Key Value pair using the KV
utility class. Do a GroupByKey to create batches of the same key type
to make the call against the database. In the DoFn, make a call out to
the database for that key and then apply the value to all values by
walking through the iterable. Follow best practices with client
instantiation as described in "Calling external services for data
enrichment".
Other relevant patterns are described in Guide to common Cloud Dataflow use-case patterns, Part 1:
Pattern: Slowly-changing lookup cache
Pattern: Calling external services for data enrichment

Kafka streams app seperate reads from writes

I am pretty new to Kafka and Kafka Streams so please bear with me. I would like to know if I am on the right track here.
I am writing to a Kafka topic at the moment and try to access the data through a rest service. The raw data kind of needs to be transformed before it will be accessed.
What I have so far is a producer that writes the raw data into a topic.
1.) Now I want streams App (should be a jar running in a container) that just transforms the data in my desired shape. Following the materialized view paradigm here.
Over simplified version of 1.)
KStreamBuilder builder = new KStreamBuilder();
KStream<String, String> source =
builder.stream("my-raw-data-topic");
KafkaStreams streams = new KafkaStreams(builder,props);
KTable<String, Long> t = source.groupByKey().count("My-Table");
streams.start();
2.) And another streams App (should be a jar running in a container) that justs holds the KTable as some sort of Repository which can be accessed via a wrapping rest service.
Here I am kind of stuck with the proper way to work with the api.
What is the bare minimun to access and query a KTable? Do I need to assign the transformation topology to the builder again?
KStreamBuilder builder = new KStreamBuilder();
KTable table = builder.table("My-Table"); //Casting?
KafkaStreams streams = new KafkaStreams(builder, props);
RestService service = new RestService(table);
// Use the Table as Repository which is wrapped by a Rest-Service and gets updated reactivly
Right now this is pseudo code
Am I on the right path here? Does is make sense to separate 1.) and 2.)? Is this the indented way to work with streams to materialize views? For me, it would have the benefit to scale up the writes and the reads via container independently where I see more traffic.
How is the repopulating of the KTable handled on a crash of either 1.) or 2.). Is this done via replication to the streaming api or is this something I would need to address via code. Like resetting the cursor and reply the events?

Couple of comments:
In your code snippet (1) you modify your topology after you handed the builder into the KafkaStreams constructor:
KafkaStreams streams = new KafkaStreams(builder,props);
// don't modify builder anymore!
You should not do this but first specify you topology and afterwards create the KafkaStreams instance.
About splitting you application into two. This can make sense to scale both parts independently. But it's hard to say in general. However, if you do spit both, the first one needs to write the transformed date into an output topic and the second one should read this output topic as a table (builder.table("output-topic-of-transformation") to serve the REST requests.
For accessing the store of the KTable, you need to get a query handle via the provided store name:
ReadOnlyKeyValueStore keyValueStore =
streams.store("My-Table", QueryableStoreTypes.keyValueStore());
See the docs for further details:
http://docs.confluent.io/current/streams/developer-guide.html#interactive-queries

Apache Kafka order windowed messages based on their value

I'm trying to find a way to re-order messages within a topic partition and send ordered messages to a new topic.
I have Kafka publisher that sends String messages of the following format:
{system_timestamp}-{event_name}?{parameters}
for example:
1494002667893-client.message?chatName=1c&messageBody=hello
1494002656558-chat.started?chatName=1c&chatPatricipants=3
Also, we add some message key for each message, to send them to the corresponding partition.
What I want to do is reorder events based on {system-timestamp} part of the message and within a 1-minute window, cause our publishers doesn't guarantee that messages will be sent in accordance with {system-timestamp} value.
For example, we can deliver to the topic, a message with a bigger {system-timestamp} value first.
I've investigated Kafka Stream API and found some examples regarding messages windowing and aggregation:
Properties streamsConfiguration = new Properties();
streamsConfiguration.put(StreamsConfig.APPLICATION_ID_CONFIG, "stream-sorter");
streamsConfiguration.put(StreamsConfig.BOOTSTRAP_SERVERS_CONFIG, "localhost:9092");
streamsConfiguration.put(StreamsConfig.ZOOKEEPER_CONNECT_CONFIG, "localhost:2181");
streamsConfiguration.put(StreamsConfig.KEY_SERDE_CLASS_CONFIG, Serdes.String().getClass().getName());
streamsConfiguration.put(StreamsConfig.VALUE_SERDE_CLASS_CONFIG, Serdes.String().getClass().getName());
KStreamBuilder builder = new KStreamBuilder();
KStream<String, String> stream = builder.stream("events");
KGroupedStream<String>, String> groupedStream = stream.groupByKey();//grouped events within partion.
/* commented since I think that I don't need any aggregation, but I guess without aggregation I can't use time windowing.
KTable<Windowed<String>, String> windowedEvents = stream.groupByKey().aggregate(
() -> "", // initial value
(aggKey, value, aggregate) -> aggregate + "", // aggregating value
TimeWindows.of(1000), // intervals in milliseconds
Serdes.String(), // serde for aggregated value
"test-store"
);*/
But what should I do next with this grouped stream? I don't see any 'sort() (e1,e2) -> e1.compareTo(e2)' methods available, also windows could be applied to methods like aggregation(), reduce() ,count() , but I think that I don't need any messages data manipulations.
How can I re-order message in the 1-minute window and send them to another topic?

Here's an outline:
Create a Processor implementation that:
in process() method, for each message:
reads the timestamp from the message value
inserts into a KeyValueStore using (timestamp, message-key) pair as the key and the message-value as the value. NB this also provides de-duplication. You'll need to provide a custom Serde to serialize the key so that the timestamp comes first, byte-wise, so that ranged queries are ordered by timestamp first.
in the punctuate() method:
reads the store using a ranged fetch from 0 to timestamp - 60'000 (=1 minute)
sends the fetched messages in order using context.forward() and deletes them from the store
The problem with this approach is that punctuate() is not triggered if no new msgs arrive to advance the "stream time". If this is a risk in your case, you can create an external scheduler that sends periodic "tick" messages to each(!) partition of your topic, that your processor should just ignore, but they'll cause punctuate to trigger in the absence of "real" msgs.
KIP-138 will address this limitation by adding explicit support for system-time punctuation:
https://cwiki.apache.org/confluence/display/KAFKA/KIP-138%3A+Change+punctuate+semantics

Here is how I ordered streams in my project.
Created topology with source, processor, sink.
In Processor
process(key, value) -> Added each record to List(instance variable).
Init() -> schedule(WINDOW_BUFFER_TIME, WALL_CLOCK_TIME) -> punctuate (timestamp) sort list of items of window buffer time in List (instance variable) and iterate and forward. Clear List (instance variable).
This logic is working fine for me.

Reading documents from the Couchbase bucket as batches

I have a Couchbase cluster which has around 25M documents. I am able to read them sequentially and also I have a function that can read a specific number of documents from the database. But my use case is slightly different since I cannot store all the 25M documents (each document is huge) in memory.
I need to process the documents in batches, say 1M/batch, push that batch to my memory, (do some operation on those documents) and push the next batch.
The function which I have written to read specific number of documents doesn't ensure that it returns a different set of documents when called again.
Is there a way by which I can complete this functionality? I also have a function which can create documents in batches. I am not sure if I can write a similar function that can read the documents in batches. The function is given below.
public void createMultipleCustomerDocuments(String docId, Customer myCust, long numDocs) {
Gson gson = new GsonBuilder().create();
JsonObject content = JsonObject.fromJson(gson.toJson(myCust));
JsonDocument document = JsonDocument.create(docId, content);
jsonDocuments.add(document);
documentCounter++;
if (documentCounter == numDocs) {
Observable.from(jsonDocuments).flatMap(new Func1<JsonDocument, Observable<JsonDocument>>() {
public Observable<JsonDocument > call(final JsonDocument docToInsert) {
return (theBucket.async().upsert(docToInsert));
}
}).last().toBlocking().single();
documentCounter = 0;
//System.out.println("Batch counter: " + batchCounter++);
}
Can someone please help me with this?

I would try to create a view which containing all of the documents, and then querying the view with skip and limit. (Can use .startKey() and startKeyId() functions instead of skip() to avoid overhead.)
but, remember not to keep that view in a production env, it will be cpu hog.
Another option, use the DCP protocol to replicate the database into your app. but it is more work.