diff --git a/R/pkg/R/serialize.R b/R/pkg/R/serialize.R index 2081786e6f83397f1c8565f24ac22d705be87523..3169d7968f8fec6f8825401030988a1cba15f3f9 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 30edfc8a7bd9422d2934edff5a7b5e29cc3c41e8..8946348ef801c00df38b081366fbc8c0f0d69a59 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"))