Repository: spark Updated Branches: refs/heads/branch-1.6 06da5fdde -> 5601d8fd0
[SPARK-11781][SPARKR] SparkR has problem in inferring type of raw type. Author: Sun Rui <rui....@intel.com> Closes #9769 from sun-rui/SPARK-11781. (cherry picked from commit cc7a1bc9370b163f51230e5ca4be612d133a5086) Signed-off-by: Shivaram Venkataraman <shiva...@cs.berkeley.edu> Project: http://git-wip-us.apache.org/repos/asf/spark/repo Commit: http://git-wip-us.apache.org/repos/asf/spark/commit/5601d8fd Tree: http://git-wip-us.apache.org/repos/asf/spark/tree/5601d8fd Diff: http://git-wip-us.apache.org/repos/asf/spark/diff/5601d8fd Branch: refs/heads/branch-1.6 Commit: 5601d8fd0a89fb40369350db655f83d36f9d5b44 Parents: 06da5fd Author: Sun Rui <rui....@intel.com> Authored: Sun Nov 29 11:08:26 2015 -0800 Committer: Shivaram Venkataraman <shiva...@cs.berkeley.edu> Committed: Sun Nov 29 11:08:35 2015 -0800 ---------------------------------------------------------------------- R/pkg/R/DataFrame.R | 34 ++++++++++++++++++-------------- R/pkg/R/SQLContext.R | 2 +- R/pkg/R/types.R | 37 ++++++++++++++++++++--------------- R/pkg/inst/tests/test_sparkSQL.R | 6 ++++++ 4 files changed, 47 insertions(+), 32 deletions(-) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/spark/blob/5601d8fd/R/pkg/R/DataFrame.R ---------------------------------------------------------------------- diff --git a/R/pkg/R/DataFrame.R b/R/pkg/R/DataFrame.R index f89e268..a82ded9 100644 --- a/R/pkg/R/DataFrame.R +++ b/R/pkg/R/DataFrame.R @@ -793,8 +793,8 @@ setMethod("dim", setMethod("collect", signature(x = "DataFrame"), function(x, stringsAsFactors = FALSE) { - names <- columns(x) - ncol <- length(names) + dtypes <- dtypes(x) + ncol <- length(dtypes) if (ncol <= 0) { # empty data.frame with 0 columns and 0 rows data.frame() @@ -817,25 +817,29 @@ setMethod("collect", # data of complex type can be held. But getting a cell from a column # of list type returns a list instead of a vector. So for columns of # non-complex type, append them as vector. + # + # For columns of complex type, be careful to access them. + # Get a column of complex type returns a list. + # Get a cell from a column of complex type returns a list instead of a vector. col <- listCols[[colIndex]] + colName <- dtypes[[colIndex]][[1]] if (length(col) <= 0) { - df[[names[colIndex]]] <- col + df[[colName]] <- col } else { - # TODO: more robust check on column of primitive types - vec <- do.call(c, col) - if (class(vec) != "list") { - df[[names[colIndex]]] <- vec + colType <- dtypes[[colIndex]][[2]] + # Note that "binary" columns behave like complex types. + if (!is.null(PRIMITIVE_TYPES[[colType]]) && colType != "binary") { + vec <- do.call(c, col) + stopifnot(class(vec) != "list") + df[[colName]] <- vec } else { - # For columns of complex type, be careful to access them. - # Get a column of complex type returns a list. - # Get a cell from a column of complex type returns a list instead of a vector. - df[[names[colIndex]]] <- col - } + df[[colName]] <- col + } + } } + df } - df - } - }) + }) #' Limit #' http://git-wip-us.apache.org/repos/asf/spark/blob/5601d8fd/R/pkg/R/SQLContext.R ---------------------------------------------------------------------- diff --git a/R/pkg/R/SQLContext.R b/R/pkg/R/SQLContext.R index a62b25f..85541c8 100644 --- a/R/pkg/R/SQLContext.R +++ b/R/pkg/R/SQLContext.R @@ -63,7 +63,7 @@ infer_type <- function(x) { }) type <- Reduce(paste0, type) type <- paste0("struct<", substr(type, 1, nchar(type) - 1), ">") - } else if (length(x) > 1) { + } else if (length(x) > 1 && type != "binary") { paste0("array<", infer_type(x[[1]]), ">") } else { type http://git-wip-us.apache.org/repos/asf/spark/blob/5601d8fd/R/pkg/R/types.R ---------------------------------------------------------------------- diff --git a/R/pkg/R/types.R b/R/pkg/R/types.R index dae4fe8..1f06af7 100644 --- a/R/pkg/R/types.R +++ b/R/pkg/R/types.R @@ -19,25 +19,30 @@ # values are equivalent R types. This is stored in an environment to allow for # more efficient look up (environments use hashmaps). PRIMITIVE_TYPES <- as.environment(list( - "byte"="integer", - "tinyint"="integer", - "smallint"="integer", - "integer"="integer", - "bigint"="numeric", - "float"="numeric", - "double"="numeric", - "decimal"="numeric", - "string"="character", - "binary"="raw", - "boolean"="logical", - "timestamp"="POSIXct", - "date"="Date")) + "tinyint" = "integer", + "smallint" = "integer", + "int" = "integer", + "bigint" = "numeric", + "float" = "numeric", + "double" = "numeric", + "decimal" = "numeric", + "string" = "character", + "binary" = "raw", + "boolean" = "logical", + "timestamp" = "POSIXct", + "date" = "Date", + # following types are not SQL types returned by dtypes(). They are listed here for usage + # by checkType() in schema.R. + # TODO: refactor checkType() in schema.R. + "byte" = "integer", + "integer" = "integer" + )) # The complex data types. These do not have any direct mapping to R's types. COMPLEX_TYPES <- list( - "map"=NA, - "array"=NA, - "struct"=NA) + "map" = NA, + "array" = NA, + "struct" = NA) # The full list of data types. DATA_TYPES <- as.environment(c(as.list(PRIMITIVE_TYPES), COMPLEX_TYPES)) http://git-wip-us.apache.org/repos/asf/spark/blob/5601d8fd/R/pkg/inst/tests/test_sparkSQL.R ---------------------------------------------------------------------- diff --git a/R/pkg/inst/tests/test_sparkSQL.R b/R/pkg/inst/tests/test_sparkSQL.R index d3b2f20..92ec820 100644 --- a/R/pkg/inst/tests/test_sparkSQL.R +++ b/R/pkg/inst/tests/test_sparkSQL.R @@ -72,6 +72,8 @@ test_that("infer types and check types", { expect_equal(infer_type(e), "map<string,integer>") expect_error(checkType("map<integer,integer>"), "Key type in a map must be string or character") + + expect_equal(infer_type(as.raw(c(1, 2, 3))), "binary") }) test_that("structType and structField", { @@ -250,6 +252,10 @@ test_that("create DataFrame from list or data.frame", { mtcarsdf <- createDataFrame(sqlContext, mtcars) expect_equivalent(collect(mtcarsdf), mtcars) + + bytes <- as.raw(c(1, 2, 3)) + df <- createDataFrame(sqlContext, list(list(bytes))) + expect_equal(collect(df)[[1]][[1]], bytes) }) test_that("create DataFrame with different data types", { --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@spark.apache.org For additional commands, e-mail: commits-h...@spark.apache.org
