Merge branch 'main' into fix_binary_op_literals

pola-rs · Aug 4, 2023 · 3a0588a · 3a0588a
2 parents 7fee29e + 39d5fcb
commit 3a0588a
Show file tree

Hide file tree

Showing 11 changed files with 64 additions and 14 deletions.
diff --git a/NEWS.md b/NEWS.md
@@ -2,7 +2,10 @@
 
 ## What's changed
 
-- New method `$explode()` for `DataFrame` and `LazyFrame`.
+- New method `$explode()` for `DataFrame` and `LazyFrame` (#314).
+- New method `$clone()` for `LazyFrame` (#347).
+- `$with_column()` is now deprecated (following upstream `polars`). It will be
+  removed in 0.9.0. It should be replaced with `$with_columns()` (#313).
 
 # polars 0.7.0
 

diff --git a/R/dataframe__frame.R b/R/dataframe__frame.R
@@ -835,6 +835,7 @@ DataFrame_with_columns = function(...) {
 #' @return DataFrame
 #' @details with_column is derived from with_columns but takes only one expression argument
 DataFrame_with_column = function(expr) {
+  warning("`with_column()` is deprecated and will be removed in polars 0.9.0. Please use `with_columns()` instead.")
   self$with_columns(expr)
 }
 

diff --git a/R/extendr-wrappers.R b/R/extendr-wrappers.R
@@ -937,7 +937,10 @@ LazyFrame$groupby <- function(exprs, maintain_order) .Call(wrap__LazyFrame__grou
 
 LazyFrame$with_columns <- function(exprs) .Call(wrap__LazyFrame__with_columns, self, exprs)
 
-LazyFrame$with_column <- function(expr) .Call(wrap__LazyFrame__with_column, self, expr)
+LazyFrame$with_column <- function(expr) {
+  warning("`with_column()` is deprecated and will be removed in polars 0.9.0. Please use `with_columns()` instead.")
+  .Call(wrap__LazyFrame__with_column, self, expr)
+}
 
 LazyFrame$with_row_count <- function(name, offset) .Call(wrap__LazyFrame__with_row_count, self, name, offset)
 
@@ -955,6 +958,8 @@ LazyFrame$schema <- function() .Call(wrap__LazyFrame__schema, self)
 
 LazyFrame$explode <- function(columns, dotdotdot_args) .Call(wrap__LazyFrame__explode, self, columns, dotdotdot_args)
 
+LazyFrame$clone_see_me_macro <- function() .Call(wrap__LazyFrame__clone_see_me_macro, self)
+
 #' @export
 `$.LazyFrame` <- function (self, name) { func <- LazyFrame[[name]]; environment(func) <- environment(); func }
 

diff --git a/R/lazyframe__lazy.R b/R/lazyframe__lazy.R
@@ -947,3 +947,14 @@ LazyFrame_explode = function(columns = list(), ...) {
   .pr$LazyFrame$explode(self, columns, dotdotdot_args) |>
     unwrap("in explode():")
 }
+
+#' Clone a LazyFrame
+#'
+#' This makes a very cheap deep copy/clone of an existing `LazyFrame`.
+#' @return A LazyFrame
+#' @examples
+#' pl$LazyFrame(mtcars)$clone()
+
+LazyFrame_clone = function() {
+  .pr$LazyFrame$clone_see_me_macro(self)
+}
diff --git a/man/LazyFrame_clone.Rd b/man/LazyFrame_clone.Rd
diff --git a/src/rust/src/lazy/dataframe.rs b/src/rust/src/lazy/dataframe.rs
@@ -401,6 +401,10 @@ impl LazyFrame {
         }
         Ok(self.0.clone().explode(columns).into())
     }
+
+    pub fn clone_see_me_macro(&self) -> LazyFrame {
+        self.clone()
+    }
 }
 
 #[derive(Clone)]

diff --git a/tests/testthat/test-expr.R b/tests/testthat/test-expr.R
@@ -2282,7 +2282,7 @@ test_that("shrink_dtype", {
     f = c("a", "b", "c"),
     g = c(0.1, 1.32, 0.12),
     h = c(T, NA, F)
-  )$with_column(pl$col("b")$cast(pl$Int64) * 32L)$select(pl$all()$shrink_dtype())
+  )$with_columns(pl$col("b")$cast(pl$Int64) * 32L)$select(pl$all()$shrink_dtype())
 
   expect_true(all(mapply(
     df$dtypes,

diff --git a/tests/testthat/test-expr_arr.R b/tests/testthat/test-expr_arr.R
@@ -1,6 +1,6 @@
 test_that("arr$lengths", {
   df = pl$DataFrame(list_of_strs = pl$Series(list(c("a", "b"), "c", character(), list(), NULL)))
-  l = df$with_column(pl$col("list_of_strs")$arr$lengths()$alias("list_of_strs_lengths"))$to_list()
+  l = df$with_columns(pl$col("list_of_strs")$arr$lengths()$alias("list_of_strs_lengths"))$to_list()
 
   expect_identical(
     l |> lapply(\(x) if (inherits(x, "integer64")) as.numeric(x) else x),
@@ -424,7 +424,7 @@ test_that("to_struct", {
 
 test_that("eval", {
   df = pl$DataFrame(a = list(a = c(1, 8, 3), b = c(4, 5, 2)))
-  l_act = df$select(pl$all()$cast(pl$dtypes$Float64))$with_column(
+  l_act = df$select(pl$all()$cast(pl$dtypes$Float64))$with_columns(
     pl$concat_list(c("a", "b"))$arr$eval(pl$element()$rank())$alias("rank")
   )$to_list()
   expect_identical(

diff --git a/tests/testthat/test-expr_binary.R b/tests/testthat/test-expr_binary.R
@@ -55,7 +55,7 @@ test_that("bin$encode and bin$decode", {
   $cast(pl$Binary)
   $alias("base64"))
 
-  test_hex_decode = encoded_hex$with_column(
+  test_hex_decode = encoded_hex$with_columns(
     pl$col("hex")$bin$decode("hex")$alias("hex_decoded")
   )$select(
     c("hex_decoded")
@@ -65,7 +65,7 @@ test_that("bin$encode and bin$decode", {
     )
   )$to_list()
 
-  test_base64_decode = encoded_base64$with_column(
+  test_base64_decode = encoded_base64$with_columns(
     pl$col("base64")$bin$decode("base64")$alias("base64_decoded")
   )$select(
     c("base64_decoded")

diff --git a/tests/testthat/test-lazy.R b/tests/testthat/test-lazy.R
@@ -646,3 +646,12 @@ test_that("with_row_count", {
   lf = pl$LazyFrame(mtcars)
   expect_identical(lf$with_row_count("idx", 42)$select(pl$col("idx"))$collect()$to_data_frame()$idx, as.double(42:(41 + nrow(mtcars))))
 })
+
+test_that("cloning", {
+  pf = pl$LazyFrame(iris)
+
+  # deep copy clone rust side object, hence not same mem address
+  pf2 = pf$clone()
+  expect_identical(pf$collect()$to_data_frame(), pf2$collect()$to_data_frame())
+  expect_different(pl$mem_address(pf), pl$mem_address(pf2))
+})
diff --git a/tests/testthat/test-lazy_functions.R b/tests/testthat/test-lazy_functions.R
@@ -23,15 +23,15 @@ test_that("pl$sum", {
   # support sum over list of expressions, wildcards or strings
   l = list(a = 1:2, b = 3:4, c = 5:6)
   expect_identical(
-    pl$DataFrame(l)$with_column(pl$sum(list("a", "c", 42L)))$to_list(),
+    pl$DataFrame(l)$with_columns(pl$sum(list("a", "c", 42L)))$to_list(),
     c(l, list(sum = c(48L, 50L)))
   )
   expect_identical(
-    pl$DataFrame(l)$with_column(pl$sum(list("*")))$to_list(),
+    pl$DataFrame(l)$with_columns(pl$sum(list("*")))$to_list(),
     c(l, list(sum = c(9L, 12L)))
   )
   expect_identical(
-    pl$DataFrame(l)$with_column(pl$sum(list(pl$col("a") + pl$col("b"), "c")))$to_list(),
+    pl$DataFrame(l)$with_columns(pl$sum(list(pl$col("a") + pl$col("b"), "c")))$to_list(),
     c(l, list(sum = c(9L, 12L)))
   )
 })
@@ -64,13 +64,13 @@ test_that("pl$min pl$max", {
 
   # support sum over list of expressions, wildcards or strings
   l = list(a = 1:2, b = 3:4, c = 5:6)
-  expect_identical(pl$DataFrame(l)$with_column(pl$min(list("a", "c", 42L)))$to_list(), c(l, list(min = c(1:2))))
-  expect_identical(pl$DataFrame(l)$with_column(pl$max(list("a", "c", 42L)))$to_list(), c(l, list(max = c(42L, 42L))))
+  expect_identical(pl$DataFrame(l)$with_columns(pl$min(list("a", "c", 42L)))$to_list(), c(l, list(min = c(1:2))))
+  expect_identical(pl$DataFrame(l)$with_columns(pl$max(list("a", "c", 42L)))$to_list(), c(l, list(max = c(42L, 42L))))
 
 
   ## TODO polars cannot handle wildcards hey wait with testing until after PR
-  # expect_identical(pl$DataFrame(l)$with_column(pl$max(list("*")))$to_list(),c(l,list(min=c(1:2))))
-  # expect_identical(pl$DataFrame(l)$with_column(pl$min(list("*")))$to_list(),c(l,list(min=c(1:2))))
+  # expect_identical(pl$DataFrame(l)$with_columns(pl$max(list("*")))$to_list(),c(l,list(min=c(1:2))))
+  # expect_identical(pl$DataFrame(l)$with_columns(pl$min(list("*")))$to_list(),c(l,list(min=c(1:2))))
 })