Skip to contents

Non-aggregate functions defined for Column.

Usage

when(condition, value)

bitwise_not(x)

bitwiseNOT(x)

create_array(x, ...)

create_map(x, ...)

expr(x)

greatest(x, ...)

input_file_name(x = "missing")

isnan(x)

least(x, ...)

lit(x)

monotonically_increasing_id(x = "missing")

nanvl(y, x)

negate(x)

rand(seed)

randn(seed)

spark_partition_id(x = "missing")

struct(x, ...)

# S4 method for class 'ANY'
lit(x)

# S4 method for class 'Column'
bitwise_not(x)

# S4 method for class 'Column'
bitwiseNOT(x)

# S4 method for class 'Column'
coalesce(x, ...)

# S4 method for class 'Column'
isnan(x)

# S4 method for class 'Column'
is.nan(x)

# S4 method for class 'missing'
monotonically_increasing_id()

# S4 method for class 'Column'
negate(x)

# S4 method for class 'missing'
spark_partition_id()

# S4 method for class 'characterOrColumn'
struct(x, ...)

# S4 method for class 'Column'
nanvl(y, x)

# S4 method for class 'Column'
greatest(x, ...)

# S4 method for class 'Column'
least(x, ...)

# S4 method for class 'character'
expr(x)

# S4 method for class 'missing'
rand(seed)

# S4 method for class 'numeric'
rand(seed)

# S4 method for class 'missing'
randn(seed)

# S4 method for class 'numeric'
randn(seed)

# S4 method for class 'Column'
when(condition, value)

# S4 method for class 'Column'
ifelse(test, yes, no)

# S4 method for class 'Column'
create_array(x, ...)

# S4 method for class 'Column'
create_map(x, ...)

# S4 method for class 'missing'
input_file_name()

Arguments

condition

the condition to test on. Must be a Column expression.

value

result expression.

x

Column to compute on. In lit, it is a literal value or a Column. In expr, it contains an expression character object to be parsed.

...

additional Columns.

y

Column to compute on.

seed

a random seed. Can be missing.

test

a Column expression that describes the condition.

yes

return values for TRUE elements of test.

no

return values for FALSE elements of test.

Details

lit: A new Column is created to represent the literal value. If the parameter is a Column, it is returned unchanged.

bitwise_not: Computes bitwise NOT.

bitwiseNOT: Computes bitwise NOT.

coalesce: Returns the first column that is not NA, or NA if all inputs are.

isnan: Returns true if the column is NaN.

is.nan: Alias for isnan.

monotonically_increasing_id: Returns a column that generates monotonically increasing 64-bit integers. The generated ID is guaranteed to be monotonically increasing and unique, but not consecutive. The current implementation puts the partition ID in the upper 31 bits, and the record number within each partition in the lower 33 bits. The assumption is that the SparkDataFrame has less than 1 billion partitions, and each partition has less than 8 billion records. As an example, consider a SparkDataFrame with two partitions, each with 3 records. This expression would return the following IDs: 0, 1, 2, 8589934592 (1L << 33), 8589934593, 8589934594. This is equivalent to the MONOTONICALLY_INCREASING_ID function in SQL. The method should be used with no argument. Note: the function is non-deterministic because its result depends on partition IDs.

negate: Unary minus, i.e. negate the expression.

spark_partition_id: Returns the partition ID as a SparkDataFrame column. Note that this is nondeterministic because it depends on data partitioning and task scheduling. This is equivalent to the SPARK_PARTITION_ID function in SQL.

struct: Creates a new struct column that composes multiple input columns.

nanvl: Returns the first column (y) if it is not NaN, or the second column (x) if the first column is NaN. Both inputs should be floating point columns (DoubleType or FloatType).

greatest: Returns the greatest value of the list of column names, skipping null values. This function takes at least 2 parameters. It will return null if all parameters are null.

least: Returns the least value of the list of column names, skipping null values. This function takes at least 2 parameters. It will return null if all parameters are null.

expr: Parses the expression string into the column that it represents, similar to SparkDataFrame.selectExpr

rand: Generates a random column with independent and identically distributed (i.i.d.) samples uniformly distributed in [0.0, 1.0). Note: the function is non-deterministic in general case.

randn: Generates a column with independent and identically distributed (i.i.d.) samples from the standard normal distribution. Note: the function is non-deterministic in general case.

when: Evaluates a list of conditions and returns one of multiple possible result expressions. For unmatched expressions null is returned.

ifelse: Evaluates a list of conditions and returns yes if the conditions are satisfied. Otherwise no is returned for unmatched conditions.

create_array: Creates a new array column. The input columns must all have the same data type.

create_map: Creates a new map column. The input columns must be grouped as key-value pairs, e.g. (key1, value1, key2, value2, ...). The key columns must all have the same data type, and can't be null. The value columns must all have the same data type.

input_file_name: Creates a string column with the input file name for a given row. The method should be used with no argument.

Note

lit since 1.5.0

bitwise_not since 3.2.0

bitwiseNOT since 1.5.0

coalesce(Column) since 2.1.1

isnan since 2.0.0

is.nan since 2.0.0

negate since 1.5.0

spark_partition_id since 2.0.0

struct since 1.6.0

nanvl since 1.5.0

greatest since 1.5.0

least since 1.5.0

expr since 1.5.0

rand since 1.5.0

rand(numeric) since 1.5.0

randn since 1.5.0

randn(numeric) since 1.5.0

when since 1.5.0

ifelse since 1.5.0

create_array since 2.3.0

create_map since 2.3.0

input_file_name since 2.3.0

See also

coalesce,SparkDataFrame-method

Other non-aggregate functions: column(), not()

Examples

if (FALSE) { # \dontrun{
# Dataframe used throughout this doc
df <- createDataFrame(cbind(model = rownames(mtcars), mtcars))} # }

if (FALSE) { # \dontrun{
tmp <- mutate(df, v1 = lit(df$mpg), v2 = lit("x"), v3 = lit("2015-01-01"),
                  v4 = negate(df$mpg), v5 = expr('length(model)'),
                  v6 = greatest(df$vs, df$am), v7 = least(df$vs, df$am),
                  v8 = column("mpg"))
head(tmp)} # }

if (FALSE) { # \dontrun{
head(select(df, bitwise_not(cast(df$vs, "int"))))} # }

if (FALSE) head(select(df, monotonically_increasing_id())) # \dontrun{}

if (FALSE) head(select(df, spark_partition_id())) # \dontrun{}

if (FALSE) { # \dontrun{
tmp <- mutate(df, v1 = struct(df$mpg, df$cyl), v2 = struct("hp", "wt", "vs"),
                  v3 = create_array(df$mpg, df$cyl, df$hp),
                  v4 = create_map(lit("x"), lit(1.0), lit("y"), lit(-1.0)))
head(tmp)} # }

if (FALSE) { # \dontrun{
tmp <- mutate(df, r1 = rand(), r2 = rand(10), r3 = randn(), r4 = randn(10))
head(tmp)} # }

if (FALSE) { # \dontrun{
tmp <- mutate(df, mpg_na = otherwise(when(df$mpg > 20, df$mpg), lit(NaN)),
                  mpg2 = ifelse(df$mpg > 20 & df$am > 0, 0, 1),
                  mpg3 = ifelse(df$mpg > 20, df$mpg, 20.0))
head(tmp)
tmp <- mutate(tmp, ind_na1 = is.nan(tmp$mpg_na), ind_na2 = isnan(tmp$mpg_na))
head(select(tmp, coalesce(tmp$mpg_na, tmp$mpg)))
head(select(tmp, nanvl(tmp$mpg_na, tmp$hp)))} # }

if (FALSE) { # \dontrun{
tmp <- read.text("README.md")
head(select(tmp, input_file_name()))} # }