diff --git a/R/pkg/R/DataFrame.R b/R/pkg/R/DataFrame.R index e12b58e2eefc50a4f257f7c621a12931b1361839..a92450274e07786ba1656295e5b0cb5f7c19fbc0 100644 --- a/R/pkg/R/DataFrame.R +++ b/R/pkg/R/DataFrame.R @@ -397,7 +397,11 @@ setMethod("coltypes", } if (is.null(type)) { - stop(paste("Unsupported data type: ", x)) + specialtype <- specialtypeshandle(x) + if (is.null(specialtype)) { + stop(paste("Unsupported data type: ", x)) + } + type <- PRIMITIVE_TYPES[[specialtype]] } } type @@ -1063,6 +1067,13 @@ setMethod("collect", df[[colIndex]] <- col } else { colType <- dtypes[[colIndex]][[2]] + if (is.null(PRIMITIVE_TYPES[[colType]])) { + specialtype <- specialtypeshandle(colType) + if (!is.null(specialtype)) { + colType <- specialtype + } + } + # Note that "binary" columns behave like complex types. if (!is.null(PRIMITIVE_TYPES[[colType]]) && colType != "binary") { vec <- do.call(c, col) diff --git a/R/pkg/R/types.R b/R/pkg/R/types.R index ad048b1cd17953b9b49a4269fb922336743fc67e..abca703617c7bcb0b77f2eded74220d2e93f00ef 100644 --- a/R/pkg/R/types.R +++ b/R/pkg/R/types.R @@ -67,3 +67,19 @@ rToSQLTypes <- as.environment(list( "double" = "double", "character" = "string", "logical" = "boolean")) + +# Helper function of coverting decimal type. When backend returns column type in the +# format of decimal(,) (e.g., decimal(10, 0)), this function coverts the column type +# as double type. This function converts backend returned types that are not the key +# of PRIMITIVE_TYPES, but should be treated as PRIMITIVE_TYPES. +# @param A type returned from the JVM backend. +# @return A type is the key of the PRIMITIVE_TYPES. +specialtypeshandle <- function(type) { + returntype <- NULL + m <- regexec("^decimal(.+)$", type) + matchedStrings <- regmatches(type, m) + if (length(matchedStrings[[1]]) >= 2) { + returntype <- "double" + } + returntype +} diff --git a/R/pkg/inst/tests/testthat/test_sparkSQL.R b/R/pkg/inst/tests/testthat/test_sparkSQL.R index 8ff56eba1f7bf56353ce7381fd381a3e5f0142fb..683a15cb4ffcdd0bfc7c5c42fe12dfc305009976 100644 --- a/R/pkg/inst/tests/testthat/test_sparkSQL.R +++ b/R/pkg/inst/tests/testthat/test_sparkSQL.R @@ -526,6 +526,17 @@ test_that( expect_is(newdf, "SparkDataFrame") expect_equal(count(newdf), 1) dropTempView("table1") + + createOrReplaceTempView(df, "dfView") + sqlCast <- collect(sql("select cast('2' as decimal) as x from dfView limit 1")) + out <- capture.output(sqlCast) + expect_true(is.data.frame(sqlCast)) + expect_equal(names(sqlCast)[1], "x") + expect_equal(nrow(sqlCast), 1) + expect_equal(ncol(sqlCast), 1) + expect_equal(out[1], " x") + expect_equal(out[2], "1 2") + dropTempView("dfView") }) test_that("test cache, uncache and clearCache", { @@ -2089,6 +2100,9 @@ test_that("Method coltypes() to get and set R's data types of a DataFrame", { # Test primitive types DF <- createDataFrame(data, schema) expect_equal(coltypes(DF), c("integer", "logical", "POSIXct")) + createOrReplaceTempView(DF, "DFView") + sqlCast <- sql("select cast('2' as decimal) as x from DFView limit 1") + expect_equal(coltypes(sqlCast), "numeric") # Test complex types x <- createDataFrame(list(list(as.environment( @@ -2132,6 +2146,14 @@ test_that("Method str()", { "setosa\" \"setosa\" \"setosa\" \"setosa\"")) expect_equal(out[7], " $ col : logi TRUE TRUE TRUE TRUE TRUE TRUE") + createOrReplaceTempView(irisDF2, "irisView") + + sqlCast <- sql("select cast('2' as decimal) as x from irisView limit 1") + castStr <- capture.output(str(sqlCast)) + expect_equal(length(castStr), 2) + expect_equal(castStr[1], "'SparkDataFrame': 1 variables:") + expect_equal(castStr[2], " $ x: num 2") + # A random dataset with many columns. This test is to check str limits # the number of columns. Therefore, it will suffice to check for the # number of returned rows