From c1961738fa366cc92d7e9a324c0e0445ca6d26b4 Mon Sep 17 00:00:00 2001
From: Yu ISHIKAWA <yuu.ishikawa@gmail.com>
Date: Tue, 23 Jun 2015 07:10:25 +0900
Subject: [PATCH 1/4] [SPARK-8431][SparkR] Add in operator to DataFrame Column
 in SparkR

---
 R/pkg/R/column.R                 | 11 +++++++++++
 R/pkg/inst/tests/test_sparkSQL.R | 10 ++++++++++
 2 files changed, 21 insertions(+)

diff --git a/R/pkg/R/column.R b/R/pkg/R/column.R
index 80e92d3105a36..2b1b03611df52 100644
--- a/R/pkg/R/column.R
+++ b/R/pkg/R/column.R
@@ -91,6 +91,16 @@ createOperator <- function(op) {
             })
 }
 
+createInOperator <- function(op) {
+  setMethod("%in%",
+            signature(x = "Column"),
+            function(x, table) {
+              table <- listToSeq(as.list(table))
+              bar <- callJMethod(x@jc, "in", table)
+              return(column(bar))
+            })
+}
+
 createColumnFunction1 <- function(name) {
   setMethod(name,
             signature(x = "Column"),
@@ -139,6 +149,7 @@ createBinaryMathfunctions <- function(name) {
 }
 
 createMethods <- function() {
+  createInOperator()
   for (op in names(operators)) {
     createOperator(op)
   }
diff --git a/R/pkg/inst/tests/test_sparkSQL.R b/R/pkg/inst/tests/test_sparkSQL.R
index 8946348ef801c..c43d4503d27ba 100644
--- a/R/pkg/inst/tests/test_sparkSQL.R
+++ b/R/pkg/inst/tests/test_sparkSQL.R
@@ -693,6 +693,16 @@ test_that("filter() on a DataFrame", {
   filtered2 <- where(df, df$name != "Michael")
   expect_true(count(filtered2) == 2)
   expect_true(collect(filtered2)$age[2] == 19)
+
+  # test suites for %in%
+  filtered3 <- filter(df, "age in (19)")
+  expect_equal(count(filtered3), 1)
+  filtered4 <- filter(df, "age in (19, 30)")
+  expect_equal(count(filtered4), 2)
+  filtered5 <- where(df, df$age %in% c(19))
+  expect_equal(count(filtered5), 1)
+  filtered6 <- where(df, df$age %in% c(19, 30))
+  expect_equal(count(filtered6), 2)
 })
 
 test_that("join() on a DataFrame", {

From 6e37936949b4f7dd25ed521f82e9c89d0684aef0 Mon Sep 17 00:00:00 2001
From: Yu ISHIKAWA <yuu.ishikawa@gmail.com>
Date: Tue, 23 Jun 2015 07:16:56 +0900
Subject: [PATCH 2/4] Modify a variable name

---
 R/pkg/R/column.R | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/R/pkg/R/column.R b/R/pkg/R/column.R
index 2b1b03611df52..41502dbf50535 100644
--- a/R/pkg/R/column.R
+++ b/R/pkg/R/column.R
@@ -96,8 +96,8 @@ createInOperator <- function(op) {
             signature(x = "Column"),
             function(x, table) {
               table <- listToSeq(as.list(table))
-              bar <- callJMethod(x@jc, "in", table)
-              return(column(bar))
+              jc <- callJMethod(x@jc, "in", table)
+              return(column(jc))
             })
 }
 

From f4309a70c8adaf7c526a4dbf57a85e6d6bf6ab4b Mon Sep 17 00:00:00 2001
From: Yu ISHIKAWA <yuu.ishikawa@gmail.com>
Date: Tue, 23 Jun 2015 07:48:02 +0900
Subject: [PATCH 3/4] Make a `setMethod` for `%in%` be independent

---
 R/pkg/R/column.R | 23 ++++++++++++-----------
 1 file changed, 12 insertions(+), 11 deletions(-)

diff --git a/R/pkg/R/column.R b/R/pkg/R/column.R
index 41502dbf50535..60d7e8da13c61 100644
--- a/R/pkg/R/column.R
+++ b/R/pkg/R/column.R
@@ -91,16 +91,6 @@ createOperator <- function(op) {
             })
 }
 
-createInOperator <- function(op) {
-  setMethod("%in%",
-            signature(x = "Column"),
-            function(x, table) {
-              table <- listToSeq(as.list(table))
-              jc <- callJMethod(x@jc, "in", table)
-              return(column(jc))
-            })
-}
-
 createColumnFunction1 <- function(name) {
   setMethod(name,
             signature(x = "Column"),
@@ -149,7 +139,6 @@ createBinaryMathfunctions <- function(name) {
 }
 
 createMethods <- function() {
-  createInOperator()
   for (op in names(operators)) {
     createOperator(op)
   }
@@ -221,6 +210,18 @@ setMethod("cast",
             }
           })
 
+#' Specify multiple values
+#'
+#' @rdname column
+#' @return a matched value as a result of comparing with given values.
+setMethod("%in%",
+          signature(x = "Column"),
+          function(x, table) {
+            table <- listToSeq(as.list(table))
+            jc <- callJMethod(x@jc, "in", table)
+            return(column(jc))
+          })
+
 #' Approx Count Distinct
 #'
 #' @rdname column

From 1f644233cf933083cab2b209436c0cdca77a7b83 Mon Sep 17 00:00:00 2001
From: Yu ISHIKAWA <yuu.ishikawa@gmail.com>
Date: Tue, 23 Jun 2015 14:07:50 +0900
Subject: [PATCH 4/4] Modify the comment

---
 R/pkg/R/column.R | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/R/pkg/R/column.R b/R/pkg/R/column.R
index 60d7e8da13c61..8e4b0f5bf1c4d 100644
--- a/R/pkg/R/column.R
+++ b/R/pkg/R/column.R
@@ -210,10 +210,14 @@ setMethod("cast",
             }
           })
 
-#' Specify multiple values
+#' Match a column with given values.
 #'
 #' @rdname column
-#' @return a matched value as a result of comparing with given values.
+#' @return a matched values as a result of comparing with given values.
+#' \dontrun{
+#'   filter(df, "age in (10, 30)")
+#'   where(df, df$age %in% c(10, 30))
+#' }
 setMethod("%in%",
           signature(x = "Column"),
           function(x, table) {