From c1de67066fa893f1864d276c29da50311839146a Mon Sep 17 00:00:00 2001 From: Shivaram Venkataraman Date: Fri, 8 May 2015 01:08:10 -0700 Subject: [PATCH] Revert workaround in SparkR to retain grouped cols Based on reverting code added in commit https://github.com/amplab-extras/spark/commit/9a6be746efc9fafad88122fa2267862ef87aa0e1 --- R/pkg/R/group.R | 4 +--- .../scala/org/apache/spark/sql/api/r/SQLUtils.scala | 11 ----------- 2 files changed, 1 insertion(+), 14 deletions(-) diff --git a/R/pkg/R/group.R b/R/pkg/R/group.R index 02237b3672d6b..0f8fd44fdf9b2 100644 --- a/R/pkg/R/group.R +++ b/R/pkg/R/group.R @@ -103,9 +103,7 @@ setMethod("agg", } } jcols <- lapply(cols, function(c) { c@jc }) - # the GroupedData.agg(col, cols*) API does not contain grouping Column - sdf <- callJStatic("org.apache.spark.sql.api.r.SQLUtils", "aggWithGrouping", - x@sgd, listToSeq(jcols)) + sdf <- callJMethod(x@sgd, "agg", jcols[[1]], listToSeq(jcols[-1])) } else { stop("agg can only support Column or character") } diff --git a/sql/core/src/main/scala/org/apache/spark/sql/api/r/SQLUtils.scala b/sql/core/src/main/scala/org/apache/spark/sql/api/r/SQLUtils.scala index ae77f72998a22..423ecdff5804a 100644 --- a/sql/core/src/main/scala/org/apache/spark/sql/api/r/SQLUtils.scala +++ b/sql/core/src/main/scala/org/apache/spark/sql/api/r/SQLUtils.scala @@ -72,17 +72,6 @@ private[r] object SQLUtils { sqlContext.createDataFrame(rowRDD, schema) } - // A helper to include grouping columns in Agg() - def aggWithGrouping(gd: GroupedData, exprs: Column*): DataFrame = { - val aggExprs = exprs.map { col => - col.expr match { - case expr: NamedExpression => expr - case expr: Expression => Alias(expr, expr.simpleString)() - } - } - gd.toDF(aggExprs) - } - def dfToRowRDD(df: DataFrame): JavaRDD[Array[Byte]] = { df.map(r => rowToRBytes(r)) }