zhengruifeng opened a new pull request, #47904:
URL: https://github.com/apache/spark/pull/47904
### What changes were proposed in this pull request?
`StringIndexer` sort arrays in executors
### Why are the changes needed?
Existing implementation collect all arrays at driver, and then sort them
with a `ForkJoinPool`.
This PR aims to move sorting from driver to executors.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
CI and manually test:
```
import org.apache.spark.ml.feature.StringIndexer
val numCol = 300
val data = (0 to 10000).map { i =>
(i, 100 * i)
}
var df = data.toDF("id", "label0")
(1 to numCol).foreach { idx =>
df = df.withColumn(s"label$idx", col("label0") + 1)
}
val inputCols = (0 to numCol).map(i => s"label$i").toArray
val outputCols = (0 to numCol).map(i => s"labelIndex$i").toArray
df.cache()
df.count()
Seq.range(0, 10).foreach(i => new
StringIndexer().setInputCols(inputCols).setOutputCols(outputCols).setStringOrderType("alphabetDesc").fit(df))
// warm up
val t0 = System.nanoTime()
Seq.range(0, 10).foreach(i => new
StringIndexer().setInputCols(inputCols).setOutputCols(outputCols).setStringOrderType("alphabetDesc").fit(df))
val t1 = System.nanoTime()
println("Elapsed time: " + (t1 - t0) / 1000000000.0 + "s")
```
master:
```
Elapsed time: 12.181134084s
```
this PR:
```
Elapsed time: 11.432259833s
```
The performance is similar, the main change is that the sort happens in
executors.
### Was this patch authored or co-authored using generative AI tooling?
No
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
