Why clustering seems doesn't work in spark cogroup function - scala

I have two hive clustered tables t1 and t2
CREATE EXTERNAL TABLE `t1`(
`t1_req_id` string,
...
PARTITIONED BY (`t1_stats_date` string)
CLUSTERED BY (t1_req_id) INTO 1000 BUCKETS
// t2 looks similar with same amount of buckets
the code looks like as following:
val t1 = spark.table("t1").as[T1].rdd.map(v => (v.t1_req_id, v))
val t2= spark.table("t2").as[T2].rdd.map(v => (v.t2_req_id, v))
val outRdd = t1.cogroup(t2)
.flatMap { coGroupRes =>
val key = coGroupRes._1
val value: (Iterable[T1], Iterable[T2])= coGroupRes._2
val t3List = // create a list with some logic on Iterable[T1] and Iterable[T2]
t3List
}
outRdd.write....
I make sure that the both t1 and t2 table has same amount of partitions, and on spark-submit there are
spark.sql.sources.bucketing.enabled=true and spark.sessionState.conf.bucketingEnabled=true flags
But Spark DAG doesn't show any impact of clustering. It seems there is still data full shuffle
What am I missing, any other configurations, tunings? How can it be assured that there is no full data shuffle?
My spark version is 2.3.1

And it shouldn't show.
Any logical optimizations are limited to DataFrame API. Once you push data to black-box functional dataset API (see Spark 2.0 Dataset vs DataFrame) and later to RDD API, no more information is pushed back to the optimizer.
You could partially utilize bucketing by performing join first, getting something around these lines
spark.table("t1")
.join(spark.table("t2"), $"t1.t1_req_id" === $"t2.t2_req_id", "outer")
.groupBy($"t1.v.t1_req_id", $"t2.t2_req_id")
.agg(...) // For example collect_set($"t1.v"), collect_set($"t2.v")
However, unlike cogroup, this will generate full Cartesian Products within groups, and might be not applicable in your case

Related

Joining two clustered tables in spark dataset seems to end up with full shuffle

I have two hive clustered tables t1 and t2
CREATE EXTERNAL TABLE `t1`(
`t1_req_id` string,
...
PARTITIONED BY (`t1_stats_date` string)
CLUSTERED BY (t1_req_id) INTO 1000 BUCKETS
// t2 looks similar with same amount of buckets
The insert part happens in hive
set hive.exec.dynamic.partition=true;
set hive.exec.dynamic.partition.mode=nonstrict;
insert overwrite table `t1` partition(t1_stats_date,t1_stats_hour)
select *
from t1_raw
where t1_stats_date='2020-05-10' and t1_stats_hour='12' AND
t1_req_id is not null
The code looks like as following:
val t1 = spark.table("t1").as[T1]
val t2= spark.table("t2").as[T2]
val outDS = t1.joinWith(t2, t1("t1_req_id) === t2("t2_req_id), "fullouter")
.map { case (t1Obj, t2Obj) =>
val t3:T3 = // do some logic
t3
}
outDS.toDF.write....
I see projection in DAG - but it seems that the job still does full data shuffle
Also, while looking into the logs of executor I don't see it reads the same bucket of the two tables in one chunk - that what I would expect to find
There are spark.sql.sources.bucketing.enabled, spark.sessionState.conf.bucketingEnabled and
spark.sql.join.preferSortMergeJoin flags
What am I missing? and why is there still full shuffle, if there are bucketed tables?
The current spark version is 2.3.1
One possibility here to check for is if you have a type mismatch. E.g. if the type of the join column is string in T1 and BIGINT in T2. Even if the types are both integer (e.g. one is INT, another BIGINT) Spark will still add shuffle here because different types use different hash functions for bucketing.

Spark Enhance Join between Terabytes of Datasets

I have five Hive tables assume the names is A, B, C, D, and E. For each table there is a customer_id as the key for join between them. Also, Each table contains at least 100:600 columns all of them is Parquet format.
Example of one table below:
CREATE TABLE table_a
(
customer_id Long,
col_1 STRING,
col_2 STRING,
col_3 STRING,
.
.
col_600 STRING
)
STORED AS PARQUET;
I need to achieve two points,
Join all of them together with the most optimum way using Spark Scala. I tried to sortByKey before join but still there is a performance bottleneck. I tried to reparation by key before join but the performance is still not good. I tried to increase the parallelism for Spark to make it 6000 with many executors but not able to achieve a good results.
After join I need to apply a separate function for some of these columns.
Sample of the join I tried below,
val dsA = spark.table(table_a)
val dsB = spark.table(table_b)
val dsC = spark.table(table_c)
val dsD = spark.table(table_d)
val dsE = spark.table(table_e)
val dsAJoineddsB = dsA.join(dsB,Seq(customer_id),"inner")
I think in this case the direct join is not the optimal case. You can acheive this task using the below simple way.
First, create case class for example FeatureData with two fields case class FeatureData(customer_id:Long, featureValue:Map[String,String])
Second, You will map each table to FeatureData case class key, [feature_name,feature_value]
Third, You will groupByKey and union all the dataset with the same key.
I the above way it will be faster to union than join. But it need more work.
After that, you will have a dataset with key,map. You will apply the transformation for key, Map(feature_name).
Simple example of the implementation as following:
You will map first the dataset to the case class then you can union all of them. After that you will groupByKey then map it and reduce it.
case class FeatureMappedData(customer_id:Long, feature: Map[String, String])
val dsAMapped = dsA.map(row ⇒
FeatureMappedData(row.customer_id,
Map("featureA" -> row.featureA,
"featureB" -> row.featureB)))
val unionDataSet = dsAMapped union dsBMapped
unionDataSet.groupByKey(_.customer_id)
.mapGroups({
case (eid, featureIter) ⇒ {
val featuresMapped: Map[String, String] = featureIter.map(_.feature).reduce(_ ++ _).withDefaultValue("0")
FeatureMappedData(customer_id, featuresMapped)
}
})

Apache Spark SQL: how to optimize chained join for dataframe

I have to make a left join between a principle data frame and several reference data frame, so a chained join computation. And I wonder how to make this action efficient and scalable.
Method 1 is easy to understand, which is also the current method, but I'm not satisfied because all the transformations have been chained and waited for the final action to trigger the computation, if I continue to add transformation and the volume of data, spark will fail at the end, so this method is not scalable.
Method 1:
def pipeline(refDF1: DataFrame, refDF2: DataFrame, refDF3: DataFrame, refDF4: DataFrame, refDF5: DataFrame): DataFrame = {
val transformations: List[DataFrame => DataFrame] = List(
castColumnsFromStringToLong(ColumnsToCastToLong),
castColumnsFromStringToFloat(ColumnsToCastToFloat),
renameColumns(RenameMapping),
filterAndDropColumns,
joinRefDF1(refDF1),
joinRefDF2(refDF2),
joinRefDF3(refDF3),
joinRefDF4(refDF4),
joinRefDF5(refDF5),
calculate()
)
transformations.reduce(_ andThen _)
}
pipeline(refDF1, refDF2, refDF3, refDF4, refDF5)(principleDF)
Method 2: I've not found a real way to achieve my idea, but I hope to trigger the computation of each join immediately.
according to my test, count() is too heavy for spark and useless for my application, but I don't know how to trigger the join computation with an efficient action. This kind of action is, in fact, the answer to this question.
val joinedDF_1 = castColumnsFromStringToLong(principleDF, ColumnsToCastToLong)
joinedDF_1.cache() // joinedDF is not always used multiple times, but for some data frame, it is, so I add cache() to indicate the usage
joinedDF_1.count()
val joinedDF_2 = castColumnsFromStringToFloat(joinedDF_1, ColumnsToCastToFloat)
joinedDF_2.cache()
joinedDF_2.count()
val joinedDF_3 = renameColumns(joinedDF_2, RenameMapping)
joinedDF_3.cache()
joinedDF_3.count()
val joinedDF_4 = filterAndDropColumns(joinedDF_4)
joinedDF_4.cache()
joinedDF_4.count()
...
When you want to force the computation of a given join (or any transformation that is not final) in Spark, you can use a simple show or count on your DataFrame. This kind of terminal points will force the computation of the result because otherwise it is simply not possible to execute the action.
Only after this will your DataFrame be effectively stored in your cache.
Once you're finished with a given DataFrame, don't hesitate to unpersist. This will unpersist your data if your cluster need more room for further computation.
You need to repartitions your dataset with the columns before calling the join transformation.
Example:
df1=df1.repartion(col("col1"),col("col2"))
df2=df2.repartion(col("col1"),col("col2"))
joinDF = df1.join(jf2,df1.col("col1").equals(df2.col("col1")) &....)
Try creating a new dataframe based on it.
Ex:
val dfTest = session.createDataFrame(df.rdd, df.schema).cache()
dfTest .storageLevel.useMemory // result should be a true.

How to get the row top 1 in Spark Structured Streaming?

I have an issue with Spark Streaming (Spark 2.2.1). I am developing a real time pipeline where first I get data from Kafka, second join the result with another table, then send the Dataframe to a ALS model (Spark ML) and it return a streaming Dataframe with one additional column predit. The problem is when I tried to get the row with the highest score, I couldn't find a way to resolve it.
I tried:
Apply SQL functions like Limit, Take, sort
dense_rank() function
search in StackOverflow
I read Unsupported Operations but doesn't seem to be much there.
Additional with the highest score I would send to a Kafka queue
My code is as follows:
val result = lines.selectExpr("CAST(value AS STRING)")
.select(from_json($"value", mySchema).as("data"))
//.select("data.*")
.selectExpr("cast(data.largo as int) as largo","cast(data.stock as int) as stock","data.verificavalormax","data.codbc","data.ide","data.timestamp_cli","data.tef_cli","data.nombre","data.descripcion","data.porcentaje","data.fechainicio","data.fechafin","data.descripcioncompleta","data.direccion","data.coordenadax","data.coordenaday","data.razon_social","data.segmento_app","data.categoria","data.subcategoria")
result.printSchema()
val model = ALSModel.load("ALSParaTiDos")
val fullPredictions = model.transform(result)
//fullPredictions is a streaming dataframe with a extra column "prediction", here i need the code to get the first row
val query = fullPredictions.writeStream.format("console").outputMode(OutputMode.Append()).option("truncate", "false").start()
query.awaitTermination()
Update
Maybe I was not clear, so I'm attaching an image with my problem. Also I wrote a more simple code to complement it: https://gist.github.com/.../9193c8a983c9007e8a1b6ec280d8df25
detailing what i need. Please I will appreciate any help :)
TL;DR Use stream-stream inner joins (Spark 2.3.0) or use memory sink (or a Hive table) for a temporary storage.
I think that the following sentence describes your case very well:
The problem is when I tried to get the row with the highest score, I couldn't find a way to resolve it.
Machine learning aside as it gives you a streaming Dataset with predictions so focusing on finding a maximum value in a column in a streaming Dataset is the real case here.
The first step is to calculate the max value as follows (copied directly from your code):
streaming.groupBy("idCustomer").agg(max("score") as "maxscore")
With that, you have two streaming Datasets that you can join as of Spark 2.3.0 (that has been released few days ago):
In Spark 2.3, we have added support for stream-stream joins, that is, you can join two streaming Datasets/DataFrames.
Inner joins on any kind of columns along with any kind of join conditions are supported.
Inner join the streaming Datasets and you're done.
Try this:
Implement a function that extract the max value of the column and then filter your dataframe with the max
def getDataFrameMaxRow(df:DataFrame , col:String):DataFrame = {
// get the maximum value
val list_prediction = df.select(col).toJSON.rdd
.collect()
.toList
.map { x => gson.fromJson[JsonObject](x, classOf[JsonObject])}
.map { x => x.get(col).getAsString.toInt}
val max = getMaxFromList(list_prediction)
// filter dataframe by the maximum value
val df_filtered = df.filter(df(col) === max.toString())
return df_filtered
}
def getMaxFromList(xs: List[Int]): Int = xs match {
case List(x: Int) => x
case x :: y :: rest => getMaxFromList( (if (x > y) x else y) :: rest )
}
And in the body of your code add:
import com.google.gson.JsonObject
import com.google.gson.Gson
import org.apache.spark.sql.DataFrame
val fullPredictions = model.transform(result)
val df_with_row_max = getDataFrameMaxRow(fullPredictions, "prediction")
Good Luck !!

How to tune mapping/filtering on big datasets (cross joined from two datasets)?

Spark 2.2.0
I have the following code converted from SQL script. It has been running for two hours and it's still running. Even slower than SQL Server. Is anything not done correctly?
The following is the plan,
Push table2 to all executors
Partition table1 and distribute the partitions to executors.
And each row in table2/t2 joins (cross join) each partition of table1.
So the calculation on the result of the cross-join can be run distributed/parallelly. (I wanted to, for example suppose​ I have 16 executors, keep a copy of t2 on all the 16 executors. Then divide table 1 into 16 partitions, one for each executor. Then each executor do the calculation on one partition of table 1 and t2.)
case class Cols (Id: Int, F2: String, F3: BigDecimal, F4: Date, F5: String,
F6: String, F7: BigDecimal, F8: String, F9: String, F10: String )
case class Result (Id1: Int, ID2: Int, Point: Int)
def getDataFromDB(source: String) = {
import sqlContext.sparkSession.implicits._
sqlContext.read.format("jdbc").options(Map(
"driver" -> "com.microsoft.sqlserver.jdbc.SQLServerDriver",
"url" -> jdbcSqlConn,
"dbtable" -> s"$source"
)).load()
.select("Id", "F2", "F3", "F4", "F5", "F6", "F7", "F8", "F9", "F10")
.as[Cols]
}
val sc = new SparkContext(conf)
val table1:DataSet[Cols] = getDataFromDB("table1").repartition(32).cache()
println(table1.count()) // about 300K rows
val table2:DataSet[Cols] = getDataFromDB("table2") // ~20K rows
table2.take(1)
println(table2.count())
val t2 = sc.broadcast(table2)
import org.apache.spark.sql.{functions => func}
val j = table1.joinWith(t2.value, func.lit(true))
j.map(x => {
val (l, r) = x
Result(l.Id, r.Id,
(if (l.F1!= null && r.F1!= null && l.F1== r.F1) 3 else 0)
+(if (l.F2!= null && r.F2!= null && l.F2== r.F2) 2 else 0)
+ ..... // All kind of the similiar expression
+(if (l.F8!= null && r.F8!= null && l.F8== r.F8) 1 else 0)
)
}).filter(x => x.Value >= 10)
println("Total count %d", j.count()) // This takes forever, the count will be about 100
How to rewrite it with Spark idiomatic way?
Ref: https://forums.databricks.com/questions/6747/how-do-i-get-a-cartesian-product-of-a-huge-dataset.html
(Somehow I feel as if I have seen the code already)
The code is slow because you use just a single task to load the entire dataset from the database using JDBC and despite cache it does not benefit from it.
Start by checking out the physical plan and Executors tab in web UI to find out about the single executor and the single task to do the work.
You should use one of the following to fine-tune the number of tasks for loading:
Use partitionColumn, lowerBound, upperBound options for the JDBC data source
Use predicates option
See JDBC To Other Databases in Spark's official documentation.
After you're fine with the loading, you should work on improving the last count action and add...another count action right after the following line:
val table1: DataSet[Cols] = getDataFromDB("table1").repartition(32).cache()
// trigger caching as it's lazy in Dataset API
table1.count
The reason why the entire query is slow is that you only mark table1 to be cached when an action gets executed which is exactly at the end (!) In other words, cache does nothing useful and more importantly makes the query performance even worse.
Performance will increase after you table2.cache.count too.
If you want to do cross join, use crossJoin operator.
crossJoin(right: Dataset[_]): DataFrame Explicit cartesian join with another DataFrame.
Please note the note from the scaladoc of crossJoin (no pun intended).
Cartesian joins are very expensive without an extra filter that can be pushed down.
The following requirement is already handled by Spark given all the optimizations available.
So the calculation on the result of the cross-join can be run distributed/parallelly.
That's Spark's job (again, no pun intended).
The following requirement begs for broadcast.
I wanted to, for example suppose​ I have 16 executors, keep a copy of t2 on all the 16 executors. Then divide table 1 into 16 partitions, one for each executor. Then each executor do the calculation on one partition of table 1 and t2.)
Use broadcast function to hint Spark SQL's engine to use table2 in broadcast mode.
broadcast[T](df: Dataset[T]): Dataset[T] Marks a DataFrame as small enough for use in broadcast joins.