Series$mean median std var

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: polars
 Title: Polars ported to R
-Version: 0.5.0.9000
+Version: 0.5.0.9001
 Depends: R (>= 4.1.0)
 Imports: utils, codetools
 Authors@R:

NEWS.md

@@ -2,6 +2,7 @@
 
 ## What's changed
 - `DataFrame` objects can be subsetted using brackets like standard R data frames: `pl$DataFrame(mtcars)[2:4, c("mpg", "hp")]` (#140 @vincentarelbundock)
+- `Series` gains new methods: `$mean`, `$median`, `$std`, `$var` (#170 vincentarelbundock)
 
 # polars v0.5.0
 

R/extendr-wrappers.R

@@ -1,7 +1,4 @@
 # Generated by extendr: Do not edit by hand
-
-# nolint start
-
 #
 # This file was created with the following call:
 #   .Call("wrap__make_polars_wrappers", use_symbols = TRUE, package_name = "polars")
@@ -975,12 +972,20 @@ Series$append_mut <- function(other) .Call(wrap__Series__append_mut, self, other
 
 Series$apply <- function(robj, rdatatype, strict, allow_fail_eval) .Call(wrap__Series__apply, self, robj, rdatatype, strict, allow_fail_eval)
 
+Series$mean <- function() .Call(wrap__Series__mean, self)
+
+Series$median <- function() .Call(wrap__Series__median, self)
+
 Series$min <- function() .Call(wrap__Series__min, self)
 
 Series$max <- function() .Call(wrap__Series__max, self)
 
 Series$sum <- function() .Call(wrap__Series__sum, self)
 
+Series$std <- function(ddof) .Call(wrap__Series__std, self, ddof)
+
+Series$var <- function(ddof) .Call(wrap__Series__var, self, ddof)
+
 Series$ceil <- function() .Call(wrap__Series__ceil, self)
 
 Series$floor <- function() .Call(wrap__Series__floor, self)
@@ -1015,5 +1020,3 @@ PolarsBackgroundHandle$is_exhausted <- function() .Call(wrap__PolarsBackgroundHa
 #' @export
 `[[.PolarsBackgroundHandle` <- `$.PolarsBackgroundHandle`
 
-
-# nolint end

R/series__series.R

@@ -653,6 +653,32 @@ Series_cumsum = function(reverse = FALSE) {
 #' pl$Series(c(1:2,3,Inf,4,-Inf,5))$sum() # Inf-Inf is NaN
 Series_sum = "use_extendr_wrapper"
 
+#' Mean
+#' @description  Reduce Series with mean
+#' @return Series
+#' @keywords Series
+#' @details
+#' Dtypes in {Int8, UInt8, Int16, UInt16} are cast to
+#' Int64 before meanming to prevent overflow issues.
+#' @examples
+#' pl$Series(c(1:2,NA,3,5))$mean() # a NA is dropped always
+#' pl$Series(c(1:2,NA,3,NaN,4,Inf))$mean() # NaN carries / poisons
+#' pl$Series(c(1:2,3,Inf,4,-Inf,5))$mean() # Inf-Inf is NaN
+Series_mean = "use_extendr_wrapper"
+
+#' Median
+#' @description  Reduce Series with median
+#' @return Series
+#' @keywords Series
+#' @details
+#' Dtypes in {Int8, UInt8, Int16, UInt16} are cast to
+#' Int64 before medianming to prevent overflow issues.
+#' @examples
+#' pl$Series(c(1:2,NA,3,5))$median() # a NA is dropped always
+#' pl$Series(c(1:2,NA,3,NaN,4,Inf))$median() # NaN carries / poisons
+#' pl$Series(c(1:2,3,Inf,4,-Inf,5))$median() # Inf-Inf is NaN
+Series_median = "use_extendr_wrapper"
+
 #' max
 #' @description  Reduce Series with max
 #' @return Series
@@ -679,6 +705,26 @@ Series_max = "use_extendr_wrapper"
 #' pl$Series(c(1:2,3,Inf,4,-Inf,5))$min() # Inf-Inf is NaN
 Series_min = "use_extendr_wrapper"
 
+#' @title Var
+#' @description Aggregate the columns of this Series to their variance values.
+#' @keywords Series
+#' @param ddof integer Delta Degrees of Freedom: the divisor used in the calculation is N - ddof, where N represents the number of elements. By default ddof is 1.
+#' @return A new `Series` object with applied aggregation.
+#' @examples pl$Series(1:10)$var()
+Series_var = function(ddof = 1) {
+  unwrap(.pr$Series$var(self, ddof))
+}
+
+#' @title Std
+#' @description Aggregate the columns of this Series to their standard deviation.
+#' @keywords Series
+#' @param ddof integer Delta Degrees of Freedom: the divisor used in the calculation is N - ddof, where N represents the number of elements. By default ddof is 1.
+#' @return A new `Series` object with applied aggregation.
+#' @examples pl$Series(1:10)$std()
+Series_std = function(ddof = 1) {
+  unwrap(.pr$Series$std(self, ddof))
+}
+
 #' Get data type of Series
 #' @keywords Series
 #' @return DataType

src/rust/src/series.rs

@@ -9,6 +9,7 @@ use crate::handle_type;
 use crate::make_r_na_fun;
 use crate::rdatatype::RPolarsDataType;
 use crate::utils::{r_error_list, r_result_list};
+use crate::robj_to;
 
 use crate::conversion_r_to_s::robjname2series;
 use crate::conversion_s_to_r::pl_series_to_list;
@@ -453,6 +454,14 @@ impl Series {
     //     Wrap(self.series.min_as_series().get(0)).into_py(py)
     // }
 
+    pub fn mean(&self) -> Series {
+        self.0.mean_as_series().into()
+    }
+
+    pub fn median(&self) -> Series {
+        self.0.median_as_series().into()
+    }
+
     pub fn min(&self) -> Series {
         self.0.min_as_series().into()
     }
@@ -465,6 +474,14 @@ impl Series {
         self.0.sum_as_series().into()
     }
 
+    pub fn std(&self, ddof: u8) ->  Series {
+        self.0.std_as_series(ddof).into()
+    }
+
+    pub fn var(&self, ddof: u8) ->  Series {
+        self.0.var_as_series(ddof).into()
+    }
+
     pub fn ceil(&self) -> List {
         r_result_list(
             self.0

tests/testthat/test-series.R

@@ -504,3 +504,24 @@ test_that("internal method get_fmt and to_fmt_char", {
     c('"foo"', '"bar"')
   )
 })
+
+
+make_cases <- function() {
+  tibble::tribble(
+    ~ .test_name, ~ base,
+    "mean",       mean,
+    "median",     median,
+    "std",        sd,
+    "var",        var,
+  )
+}
+patrick::with_parameters_test_that("mean, median, std, var", {
+  s = pl$Series(rnorm(100))
+  a = s[[.test_name]]()
+  # upstream .std_as_series() does not appear to return Series
+  if (inherits(a, "Series")) a = a$to_vector()
+  b = base(s$to_vector())
+  expect_equal(a, b)
+  },
+  .cases = make_cases()
+)