From a5c52c1a3488b69bec19e460d2d1fdb0c9ada58d Mon Sep 17 00:00:00 2001 From: hqzizania <qian.huang@intel.com> Date: Mon, 8 Jun 2015 21:40:12 -0700 Subject: [PATCH] [SPARK-6820] [SPARKR] Convert NAs to null type in SparkR DataFrames Author: hqzizania <qian.huang@intel.com> Closes #6190 from hqzizania/R and squashes the following commits: 1641f9e [hqzizania] fixes and add test units bb3411a [hqzizania] Convert NAs to null type in SparkR DataFrames --- R/pkg/R/serialize.R | 8 +++++++ R/pkg/inst/tests/test_sparkSQL.R | 37 ++++++++++++++++++++++++++++++++ 2 files changed, 45 insertions(+) diff --git a/R/pkg/R/serialize.R b/R/pkg/R/serialize.R index 2081786e6f..3169d7968f 100644 --- a/R/pkg/R/serialize.R +++ b/R/pkg/R/serialize.R @@ -37,6 +37,14 @@ writeObject <- function(con, object, writeType = TRUE) { # passing in vectors as arrays and instead require arrays to be passed # as lists. type <- class(object)[[1]] # class of POSIXlt is c("POSIXlt", "POSIXt") + # Checking types is needed here, since ‘is.na’ only handles atomic vectors, + # lists and pairlists + if (type %in% c("integer", "character", "logical", "double", "numeric")) { + if (is.na(object)) { + object <- NULL + type <- "NULL" + } + } if (writeType) { writeType(con, type) } diff --git a/R/pkg/inst/tests/test_sparkSQL.R b/R/pkg/inst/tests/test_sparkSQL.R index 30edfc8a7b..8946348ef8 100644 --- a/R/pkg/inst/tests/test_sparkSQL.R +++ b/R/pkg/inst/tests/test_sparkSQL.R @@ -101,6 +101,43 @@ test_that("create DataFrame from RDD", { expect_equal(dtypes(df), list(c("a", "int"), c("b", "string"))) }) +test_that("convert NAs to null type in DataFrames", { + rdd <- parallelize(sc, list(list(1L, 2L), list(NA, 4L))) + df <- createDataFrame(sqlContext, rdd, list("a", "b")) + expect_true(is.na(collect(df)[2, "a"])) + expect_equal(collect(df)[2, "b"], 4L) + + l <- data.frame(x = 1L, y = c(1L, NA_integer_, 3L)) + df <- createDataFrame(sqlContext, l) + expect_equal(collect(df)[2, "x"], 1L) + expect_true(is.na(collect(df)[2, "y"])) + + rdd <- parallelize(sc, list(list(1, 2), list(NA, 4))) + df <- createDataFrame(sqlContext, rdd, list("a", "b")) + expect_true(is.na(collect(df)[2, "a"])) + expect_equal(collect(df)[2, "b"], 4) + + l <- data.frame(x = 1, y = c(1, NA_real_, 3)) + df <- createDataFrame(sqlContext, l) + expect_equal(collect(df)[2, "x"], 1) + expect_true(is.na(collect(df)[2, "y"])) + + l <- list("a", "b", NA, "d") + df <- createDataFrame(sqlContext, l) + expect_true(is.na(collect(df)[3, "_1"])) + expect_equal(collect(df)[4, "_1"], "d") + + l <- list("a", "b", NA_character_, "d") + df <- createDataFrame(sqlContext, l) + expect_true(is.na(collect(df)[3, "_1"])) + expect_equal(collect(df)[4, "_1"], "d") + + l <- list(TRUE, FALSE, NA, TRUE) + df <- createDataFrame(sqlContext, l) + expect_true(is.na(collect(df)[3, "_1"])) + expect_equal(collect(df)[4, "_1"], TRUE) +}) + test_that("toDF", { rdd <- lapply(parallelize(sc, 1:10), function(x) { list(x, as.character(x)) }) df <- toDF(rdd, list("a", "b")) -- GitLab