9dadf019b9
This PR introduces several performance optimizations to `HadoopFsRelation` and `ParquetRelation2`:
1. Moving `FileStatus` listing from `DataSourceStrategy` into a cache within `HadoopFsRelation`.
This new cache generalizes and replaces the one used in `ParquetRelation2`.
This also introduces an interface change: to reuse cached `FileStatus` objects, `HadoopFsRelation.buildScan` methods now receive `Array[FileStatus]` instead of `Array[String]`.
1. When Parquet task side metadata reading is enabled, skip reading row group information when reading Parquet footers.
This is basically what PR #5334 does. Also, now we uses `ParquetFileReader.readAllFootersInParallel` to read footers in parallel.
Another optimization in question is, instead of asking `HadoopFsRelation.buildScan` to return an `RDD[Row]` for a single selected partition and then union them all, we ask it to return an `RDD[Row]` for all selected partitions. This optimization is based on the fact that Hadoop configuration broadcasting used in `NewHadoopRDD` takes 34% time in the following microbenchmark. However, this complicates data source user code because user code must merge partition values manually.
To check the cost of broadcasting in `NewHadoopRDD`, I also did microbenchmark after removing the `broadcast` call in `NewHadoopRDD`. All results are shown below.
### Microbenchmark
#### Preparation code
Generating a partitioned table with 50k partitions, 1k rows per partition:
```scala
import sqlContext._
import sqlContext.implicits._
for (n <- 0 until 500) {
val data = for {
p <- (n * 10) until ((n + 1) * 10)
i <- 0 until 1000
} yield (i, f"val_$i%04d", f"$p%04d")
data.
toDF("a", "b", "p").
write.
partitionBy("p").
mode("append").
parquet(path)
}
```
#### Benchmarking code
```scala
import sqlContext._
import sqlContext.implicits._
import org.apache.spark.sql.types._
import com.google.common.base.Stopwatch
val path = "hdfs://localhost:9000/user/lian/5k"
def benchmark(n: Int)(f: => Unit) {
val stopwatch = new Stopwatch()
def run() = {
stopwatch.reset()
stopwatch.start()
f
stopwatch.stop()
stopwatch.elapsedMillis()
}
val records = (0 until n).map(_ => run())
(0 until n).foreach(i => println(s"Round $i: ${records(i)} ms"))
println(s"Average: ${records.sum / n.toDouble} ms")
}
benchmark(3) { read.parquet(path).explain(extended = true) }
```
#### Results
Before:
```
Round 0: 72528 ms
Round 1: 68938 ms
Round 2: 65372 ms
Average: 68946.0 ms
```
After:
```
Round 0: 59499 ms
Round 1: 53645 ms
Round 2: 53844 ms
Round 3: 49093 ms
Round 4: 50555 ms
Average: 53327.2 ms
```
Also removing Hadoop configuration broadcasting:
(Note that I was testing on a local laptop, thus network cost is pretty low.)
```
Round 0: 15806 ms
Round 1: 14394 ms
Round 2: 14699 ms
Round 3: 15334 ms
Round 4: 14123 ms
Average: 14871.2 ms
```
Author: Cheng Lian <lian@databricks.com>
Closes #6225 from liancheng/spark-7673 and squashes the following commits:
2d58a2b [Cheng Lian] Skips reading row group information when using task side metadata reading
7aa3748 [Cheng Lian] Optimizes FileStatusCache by introducing a map from parent directories to child files
ba41250 [Cheng Lian] Reuses HadoopFsRelation FileStatusCache in ParquetRelation2
3d278f7 [Cheng Lian] Fixes a bug when reading a single Parquet data file
b84612a [Cheng Lian] Fixes Scala style issue
6a08b02 [Cheng Lian] WIP: Moves file status cache into HadoopFSRelation
|
||
|---|---|---|
| .. | ||
| catalyst | ||
| core | ||
| hive | ||
| hive-thriftserver | ||
| README.md | ||
Spark SQL
This module provides support for executing relational queries expressed in either SQL or a LINQ-like Scala DSL.
Spark SQL is broken up into four subprojects:
- Catalyst (sql/catalyst) - An implementation-agnostic framework for manipulating trees of relational operators and expressions.
- Execution (sql/core) - A query planner / execution engine for translating Catalyst’s logical query plans into Spark RDDs. This component also includes a new public interface, SQLContext, that allows users to execute SQL or LINQ statements against existing RDDs and Parquet files.
- Hive Support (sql/hive) - Includes an extension of SQLContext called HiveContext that allows users to write queries using a subset of HiveQL and access data from a Hive Metastore using Hive SerDes. There are also wrappers that allows users to run queries that include Hive UDFs, UDAFs, and UDTFs.
- HiveServer and CLI support (sql/hive-thriftserver) - Includes support for the SQL CLI (bin/spark-sql) and a HiveServer2 (for JDBC/ODBC) compatible server.
Other dependencies for developers
In order to create new hive test cases (i.e. a test suite based on HiveComparisonTest),
you will need to setup your development environment based on the following instructions.
If you are working with Hive 0.12.0, you will need to set several environmental variables as follows.
export HIVE_HOME="<path to>/hive/build/dist"
export HIVE_DEV_HOME="<path to>/hive/"
export HADOOP_HOME="<path to>/hadoop-1.0.4"
If you are working with Hive 0.13.1, the following steps are needed:
- Download Hive's 0.13.1 and set
HIVE_HOMEwithexport HIVE_HOME="<path to hive>". Please do not setHIVE_DEV_HOME(See SPARK-4119). - Set
HADOOP_HOMEwithexport HADOOP_HOME="<path to hadoop>" - Download all Hive 0.13.1a jars (Hive jars actually used by Spark) from here and replace corresponding original 0.13.1 jars in
$HIVE_HOME/lib. - Download Kryo 2.21 jar (Note: 2.22 jar does not work) and Javolution 5.5.1 jar to
$HIVE_HOME/lib. - This step is optional. But, when generating golden answer files, if a Hive query fails and you find that Hive tries to talk to HDFS or you find weird runtime NPEs, set the following in your test suite...
val testTempDir = Utils.createTempDir()
// We have to use kryo to let Hive correctly serialize some plans.
sql("set hive.plan.serialization.format=kryo")
// Explicitly set fs to local fs.
sql(s"set fs.default.name=file://$testTempDir/")
// Ask Hive to run jobs in-process as a single map and reduce task.
sql("set mapred.job.tracker=local")
Using the console
An interactive scala console can be invoked by running build/sbt hive/console.
From here you can execute queries with HiveQl and manipulate DataFrame by using DSL.
catalyst$ build/sbt hive/console
[info] Starting scala interpreter...
import org.apache.spark.sql.catalyst.analysis._
import org.apache.spark.sql.catalyst.dsl._
import org.apache.spark.sql.catalyst.errors._
import org.apache.spark.sql.catalyst.expressions._
import org.apache.spark.sql.catalyst.plans.logical._
import org.apache.spark.sql.catalyst.rules._
import org.apache.spark.sql.catalyst.util._
import org.apache.spark.sql.execution
import org.apache.spark.sql.functions._
import org.apache.spark.sql.hive._
import org.apache.spark.sql.hive.test.TestHive._
import org.apache.spark.sql.types._
Type in expressions to have them evaluated.
Type :help for more information.
scala> val query = sql("SELECT * FROM (SELECT * FROM src) a")
query: org.apache.spark.sql.DataFrame = org.apache.spark.sql.DataFrame@74448eed
Query results are DataFrames and can be operated as such.
scala> query.collect()
res2: Array[org.apache.spark.sql.Row] = Array([238,val_238], [86,val_86], [311,val_311], [27,val_27]...
You can also build further queries on top of these DataFrames using the query DSL.
scala> query.where(query("key") > 30).select(avg(query("key"))).collect()
res3: Array[org.apache.spark.sql.Row] = Array([274.79025423728814])