[SPARK-10328] [SPARKR] Fix generic for na.omit

S3 function is at https://stat.ethz.ch/R-manual/R-patched/library/stats/html/na.fail.html

Author: Shivaram Venkataraman <shivaram@cs.berkeley.edu>
Author: Shivaram Venkataraman <shivaram.venkataraman@gmail.com>
Author: Yu ISHIKAWA <yuu.ishikawa@gmail.com>

Closes #8495 from shivaram/na-omit-fix.

1 files changed, 22 insertions, 1 deletions

diff --git a/R/pkg/inst/tests/test_sparkSQL.R b/R/pkg/inst/tests/test_sparkSQL.R
index 4b672e115f..933b11c8ee 100644
--- a/R/pkg/inst/tests/test_sparkSQL.R
+++ b/R/pkg/inst/tests/test_sparkSQL.R
@@ -1083,7 +1083,7 @@ test_that("describe() and summarize() on a DataFrame", {
   expect_equal(collect(stats2)[5, "age"], "30")
 })
 
-test_that("dropna() on a DataFrame", {
+test_that("dropna() and na.omit() on a DataFrame", {
   df <- jsonFile(sqlContext, jsonPathNa)
   rows <- collect(df)
 
@@ -1092,6 +1092,8 @@ test_that("dropna() on a DataFrame", {
   expected <- rows[!is.na(rows$name),]
   actual <- collect(dropna(df, cols = "name"))
   expect_identical(expected, actual)
+  actual <- collect(na.omit(df, cols = "name"))
+  expect_identical(expected, actual)
 
   expected <- rows[!is.na(rows$age),]
   actual <- collect(dropna(df, cols = "age"))
@@ -1101,48 +1103,67 @@ test_that("dropna() on a DataFrame", {
   expect_identical(expected$age, actual$age)
   expect_identical(expected$height, actual$height)
   expect_identical(expected$name, actual$name)
+  actual <- collect(na.omit(df, cols = "age"))
 
   expected <- rows[!is.na(rows$age) & !is.na(rows$height),]
   actual <- collect(dropna(df, cols = c("age", "height")))
   expect_identical(expected, actual)
+  actual <- collect(na.omit(df, cols = c("age", "height")))
+  expect_identical(expected, actual)
 
   expected <- rows[!is.na(rows$age) & !is.na(rows$height) & !is.na(rows$name),]
   actual <- collect(dropna(df))
   expect_identical(expected, actual)
+  actual <- collect(na.omit(df))
+  expect_identical(expected, actual)
 
   # drop with how
 
   expected <- rows[!is.na(rows$age) & !is.na(rows$height) & !is.na(rows$name),]
   actual <- collect(dropna(df))
   expect_identical(expected, actual)
+  actual <- collect(na.omit(df))
+  expect_identical(expected, actual)
 
   expected <- rows[!is.na(rows$age) | !is.na(rows$height) | !is.na(rows$name),]
   actual <- collect(dropna(df, "all"))
   expect_identical(expected, actual)
+  actual <- collect(na.omit(df, "all"))
+  expect_identical(expected, actual)
 
   expected <- rows[!is.na(rows$age) & !is.na(rows$height) & !is.na(rows$name),]
   actual <- collect(dropna(df, "any"))
   expect_identical(expected, actual)
+  actual <- collect(na.omit(df, "any"))
+  expect_identical(expected, actual)
 
   expected <- rows[!is.na(rows$age) & !is.na(rows$height),]
   actual <- collect(dropna(df, "any", cols = c("age", "height")))
   expect_identical(expected, actual)
+  actual <- collect(na.omit(df, "any", cols = c("age", "height")))
+  expect_identical(expected, actual)
 
   expected <- rows[!is.na(rows$age) | !is.na(rows$height),]
   actual <- collect(dropna(df, "all", cols = c("age", "height")))
   expect_identical(expected, actual)
+  actual <- collect(na.omit(df, "all", cols = c("age", "height")))
+  expect_identical(expected, actual)
 
   # drop with threshold
 
   expected <- rows[as.integer(!is.na(rows$age)) + as.integer(!is.na(rows$height)) >= 2,]
   actual <- collect(dropna(df, minNonNulls = 2, cols = c("age", "height")))
   expect_identical(expected, actual)
+  actual <- collect(na.omit(df, minNonNulls = 2, cols = c("age", "height")))
+  expect_identical(expected, actual)
 
   expected <- rows[as.integer(!is.na(rows$age)) +
                    as.integer(!is.na(rows$height)) +
                    as.integer(!is.na(rows$name)) >= 3,]
   actual <- collect(dropna(df, minNonNulls = 3, cols = c("name", "age", "height")))
   expect_identical(expected, actual)
+  actual <- collect(na.omit(df, minNonNulls = 3, cols = c("name", "age", "height")))
+  expect_identical(expected, actual)
 })
 
 test_that("fillna() on a DataFrame", {