Merge remote-tracking branch 'origin-apache/master' into SPARK-22465

R/pkg/NAMESPACE

@@ -230,6 +230,7 @@ exportMethods("%<=>%",
               "date_add",
               "date_format",
               "date_sub",
+              "date_trunc",
               "datediff",
               "dayofmonth",
               "dayofweek",

R/pkg/R/functions.R

@@ -40,10 +40,17 @@ NULL
 #'
 #' @param x Column to compute on. In \code{window}, it must be a time Column of
 #'          \code{TimestampType}.
-#' @param format For \code{to_date} and \code{to_timestamp}, it is the string to use to parse
-#'               Column \code{x} to DateType or TimestampType. For \code{trunc}, it is the string
-#'               to use to specify the truncation method. For example, "year", "yyyy", "yy" for
-#'               truncate by year, or "month", "mon", "mm" for truncate by month.
+#' @param format The format for the given dates or timestamps in Column \code{x}. See the
+#'               format used in the following methods:
+#'               \itemize{
+#'               \item \code{to_date} and \code{to_timestamp}: it is the string to use to parse
+#'                    Column \code{x} to DateType or TimestampType.
+#'               \item \code{trunc}: it is the string to use to specify the truncation method.
+#'                    For example, "year", "yyyy", "yy" for truncate by year, or "month", "mon",
+#'                    "mm" for truncate by month.
+#'               \item \code{date_trunc}: it is similar with \code{trunc}'s but additionally
+#'                    supports "day", "dd", "second", "minute", "hour", "week" and "quarter".
+#'               }
 #' @param ... additional argument(s).
 #' @name column_datetime_functions
 #' @rdname column_datetime_functions
@@ -3478,3 +3485,22 @@ setMethod("trunc",
                               x@jc, as.character(format))
             column(jc)
           })
+
+#' @details
+#' \code{date_trunc}: Returns timestamp truncated to the unit specified by the format.
+#'
+#' @rdname column_datetime_functions
+#' @aliases date_trunc date_trunc,character,Column-method
+#' @export
+#' @examples
+#'
+#' \dontrun{
+#' head(select(df, df$time, date_trunc("hour", df$time), date_trunc("minute", df$time),
+#'             date_trunc("week", df$time), date_trunc("quarter", df$time)))}
+#' @note date_trunc since 2.3.0
+setMethod("date_trunc",
+          signature(format = "character", x = "Column"),
+          function(format, x) {
+            jc <- callJStatic("org.apache.spark.sql.functions", "date_trunc", format, x@jc)
+            column(jc)
+          })

R/pkg/R/generics.R

@@ -1043,6 +1043,11 @@ setGeneric("date_format", function(y, x) { standardGeneric("date_format") })
 #' @name NULL
 setGeneric("date_sub", function(y, x) { standardGeneric("date_sub") })
 
+#' @rdname column_datetime_functions
+#' @export
+#' @name NULL
+setGeneric("date_trunc", function(format, x) { standardGeneric("date_trunc") })
+
 #' @rdname column_datetime_functions
 #' @export
 #' @name NULL

R/pkg/tests/fulltests/test_sparkSQL.R

@@ -1418,6 +1418,8 @@ test_that("column functions", {
   c22 <- not(c)
   c23 <- trunc(c, "year") + trunc(c, "yyyy") + trunc(c, "yy") +
     trunc(c, "month") + trunc(c, "mon") + trunc(c, "mm")
+  c24 <- date_trunc("hour", c) + date_trunc("minute", c) + date_trunc("week", c) +
+    date_trunc("quarter", c)
 
   # Test if base::is.nan() is exposed
   expect_equal(is.nan(c("a", "b")), c(FALSE, FALSE))
@@ -1729,6 +1731,7 @@ test_that("date functions on a DataFrame", {
   expect_gt(collect(select(df2, unix_timestamp()))[1, 1], 0)
   expect_gt(collect(select(df2, unix_timestamp(df2$b)))[1, 1], 0)
   expect_gt(collect(select(df2, unix_timestamp(lit("2015-01-01"), "yyyy-MM-dd")))[1, 1], 0)
+  expect_equal(collect(select(df2, month(date_trunc("yyyy", df2$b))))[, 1], c(1, 1))
 
   l3 <- list(list(a = 1000), list(a = -1000))
   df3 <- createDataFrame(l3)

examples/src/main/scala/org/apache/spark/examples/sql/hive/SparkHiveExample.scala

@@ -135,7 +135,7 @@ object SparkHiveExample {
     hiveTableDF.coalesce(10).write.mode(SaveMode.Overwrite)
       .partitionBy("key").parquet(hiveExternalTableLocation)
     // $example off:spark_hive$
-    
+
     spark.stop()
   }
 }