Better handling of int64

NEWS.md

@@ -16,6 +16,10 @@
     variable (#708).
 -   New methods for the `list` subnamespace: `$set_union()`, `$set_intersection()`,
     `$set_difference()`, `$set_symmetric_difference()` (#712).
+-   New option `int64_conversion` to specify how Int64 columns (that don't have
+    equivalent in base R) should be converted. This option can either be set 
+    globally with `pl$set_options()` or on a case-by-case basis, e.g with 
+    `$to_data_frame(int64_conversion =)` (#706).
 
 ## polars 0.12.2
 

R/dataframe__frame.R

@@ -845,15 +845,16 @@ DataFrame_group_by = function(..., maintain_order = pl$options$maintain_order) {
 #' Return Polars DataFrame as R data.frame
 #'
 #' @param ... Any args pased to `as.data.frame()`.
+#' @inheritParams pl_set_options
 #'
 #' @return An R data.frame
 #' @keywords DataFrame
 #' @examples
 #' df = pl$DataFrame(iris[1:3, ])
 #' df$to_data_frame()
-DataFrame_to_data_frame = function(...) {
+DataFrame_to_data_frame = function(..., int64_conversion = pl$options$int64_conversion ) {
   # do not unnest structs and mark with I to also preserve categoricals as is
-  l = lapply(self$to_list(unnest_structs = FALSE), I)
+  l = lapply(self$to_list(unnest_structs = FALSE, int64_conversion = int64_conversion), I)
 
   # similar to as.data.frame, but avoid checks, whcih would edit structs
   df = data.frame(seq_along(l[[1L]]), ...)
@@ -870,6 +871,7 @@ DataFrame_to_data_frame = function(...) {
 #'
 #' @param unnest_structs Boolean. If `TRUE` (default), then `$unnest()` is applied
 #' on any struct column.
+#' @inheritParams pl_set_options
 #'
 #' @details
 #' For simplicity reasons, this implementation relies on unnesting all structs
@@ -881,11 +883,14 @@ DataFrame_to_data_frame = function(...) {
 #' @keywords DataFrame
 #' @examples
 #' pl$DataFrame(iris)$to_list()
-DataFrame_to_list = function(unnest_structs = TRUE) {
+DataFrame_to_list = function(unnest_structs = TRUE, ..., int64_conversion = pl$options$int64_conversion ) {
   if (unnest_structs) {
-    unwrap(.pr$DataFrame$to_list(self))
+    .pr$DataFrame$to_list(self, int64_conversion ) |>
+      unwrap("in $to_list():")
   } else {
-    restruct_list(unwrap(.pr$DataFrame$to_list_tag_structs(self)))
+    .pr$DataFrame$to_list_tag_structs(self, int64_conversion ) |>
+      unwrap("in $to_list():") |>
+      restruct_list()
   }
 }
 

R/expr__expr.R

@@ -3177,17 +3177,18 @@ Expr_rep_extend = function(expr, n, rechunk = TRUE, upcast = TRUE) {
 #' Otherwise, provide a DataFrame that the Expr should be evaluated in.
 #' @param i Numeric column to extract. Default is zero (which gives the first
 #' column).
+#' @inheritParams pl_set_options
 #' @return R object
 #' @examples
 #' pl$lit(1:3)$to_r()
-Expr_to_r = function(df = NULL, i = 0) {
+Expr_to_r = function(df = NULL, i = 0, ..., int64_conversion = pl$options$int64_conversion ) {
   if (is.null(df)) {
-    pl$select(self)$to_series(i)$to_r()
+    pl$select(self)$to_series(i)$to_r(int64_conversion )
   } else {
     if (!inherits(df, c("RPolarsDataFrame"))) {
       stop("Expr_to_r: input is not NULL or a DataFrame/Lazyframe")
     }
-    df$select(self)$to_series(i)$to_r()
+    df$select(self)$to_series(i)$to_r(int64_conversion )
   }
 }
 

R/extendr-wrappers.R

@@ -161,11 +161,11 @@ RPolarsDataFrame$dtype_strings <- function() .Call(wrap__RPolarsDataFrame__dtype
 
 RPolarsDataFrame$schema <- function() .Call(wrap__RPolarsDataFrame__schema, self)
 
-RPolarsDataFrame$to_list <- function() .Call(wrap__RPolarsDataFrame__to_list, self)
+RPolarsDataFrame$to_list <- function(int64_conversion) .Call(wrap__RPolarsDataFrame__to_list, self, int64_conversion)
 
-RPolarsDataFrame$to_list_unwind <- function() .Call(wrap__RPolarsDataFrame__to_list_unwind, self)
+RPolarsDataFrame$to_list_unwind <- function(int64_conversion) .Call(wrap__RPolarsDataFrame__to_list_unwind, self, int64_conversion)
 
-RPolarsDataFrame$to_list_tag_structs <- function() .Call(wrap__RPolarsDataFrame__to_list_tag_structs, self)
+RPolarsDataFrame$to_list_tag_structs <- function(int64_conversion) .Call(wrap__RPolarsDataFrame__to_list_tag_structs, self, int64_conversion)
 
 RPolarsDataFrame$equals <- function(other) .Call(wrap__RPolarsDataFrame__equals, self, other)
 
@@ -1165,7 +1165,7 @@ RPolarsSeries$sleep <- function(millis) .Call(wrap__RPolarsSeries__sleep, self,
 
 RPolarsSeries$panic <- function() .Call(wrap__RPolarsSeries__panic, self)
 
-RPolarsSeries$to_r <- function() .Call(wrap__RPolarsSeries__to_r, self)
+RPolarsSeries$to_r <- function(int64_conversion) .Call(wrap__RPolarsSeries__to_r, self, int64_conversion)
 
 RPolarsSeries$rename_mut <- function(name) invisible(.Call(wrap__RPolarsSeries__rename_mut, self, name))
 

R/options.R

@@ -25,6 +25,11 @@ polars_optreq$debug_polars = list(must_be_bool = is_bool)
 # polars_optenv$rpool_cap # active binding for getting value, not for
 polars_optreq$rpool_cap = list() # rust-side options already check args
 
+polars_optenv$int64_conversion = "double"
+polars_optreq$int64_conversion = list(
+  acceptable_choices = function(x) !is.null(x) && x %in% c("bit64", "double", "string"),
+  bit64_is_attached = function(x) if (x == "bit64") x %in% .packages() else TRUE
+)
 
 ## END OF DEFINED OPTIONS
 
@@ -62,9 +67,15 @@ polars_optreq$rpool_cap = list() # rust-side options already check args
 #' @param no_messages Hide messages.
 #' @param rpool_cap The maximum number of R sessions that can be used to process
 #' R code in the background. See Details.
+#' @param int64_conversion  How should Int64 values be handled when converting a
+#' polars object to R?
+#'
+#' * `"double"` (default) converts the integer values to double.
+#' * `"bit64"` uses `bit64::as.integer64()` to do the conversion (requires
+#'   the package `bit64` to be attached).
+#' * `"string"` converts Int64 values to character.
 #'
 #' @rdname pl_options
-#' @docType NULL
 #'
 #' @return
 #' `pl$options` returns a named list with the value (`TRUE` or `FALSE`) of
@@ -93,7 +104,8 @@ pl_set_options = function(
     do_not_repeat_call = FALSE,
     debug_polars = FALSE,
     no_messages = FALSE,
-    rpool_cap = 4) {
+    rpool_cap = 4,
+    int64_conversion = c("bit64", "double", "string")) {
   # only modify arguments that were explicitly written in the function call
   # (otherwise calling set_options() twice in a row would reset the args
   # modified in the first call)
@@ -152,14 +164,17 @@ pl_reset_options = function() {
   assign("debug_polars", FALSE, envir = polars_optenv)
   assign("no_messages", FALSE, envir = polars_optenv)
   assign("rpool_cap", 4, envir = polars_optenv)
+  assign("int64_conversion ", "double", envir = polars_optenv)
 }
 
 
 translate_failures = \(x) {
   lookups = c(
     "must_be_scalar" = "Input must be of length one.",
     "must_be_integer" = "Input must be an integer.",
-    "must_be_bool" = "Input must be TRUE or FALSE"
+    "must_be_bool" = "Input must be TRUE or FALSE.",
+    "acceptable_choices" = "`int64_conversion ` must be one of \"float\", \"string\", \"bit64\".",
+    "bit64_is_attached" = "Package `bit64` must be attached to use `int64_conversion = \"bit64\"`."
   )
   trans = lookups[x]
   trans[is.na(trans)] = x[is.na(trans)]

R/series__series.R

@@ -240,8 +240,9 @@ Series_shape = method_as_property(function() {
 
 #' Get r vector/list
 #' @description return R list (if polars Series is list)  or vector (any other polars Series type)
-#' @name Series_to_r
-#' @rdname Series_to_r
+#'
+#' @inheritParams pl_set_options
+#'
 #' @return R list or vector
 #' @keywords Series
 #' @details
@@ -274,20 +275,21 @@ Series_shape = method_as_property(function() {
 #' series_list$to_r() # as list because Series DataType is list
 #' series_list$to_r_list() # implicit call as.list(), same as to_r() as already list
 #' series_list$to_vector() # implicit call unlist(), append into a vector
-Series_to_r = \() {
-  unwrap(.pr$Series$to_r(self), "in $to_r():")
+Series_to_r = \(int64_conversion = pl$options$int64_conversion ) {
+  unwrap(.pr$Series$to_r(self, int64_conversion ), "in $to_r():")
 }
 # TODO replace list example with Series only syntax
 
 #' @rdname Series_to_r
 #' @name Series_to_vector
 #' @description return R vector (implicit unlist)
+#' @inheritParams pl_set_options
 #' @return R vector
 #' @keywords Series
 #' series_vec = pl$Series(letters[1:3])
 #' series_vec$to_vector()
-Series_to_vector = \() {
-  unlist(unwrap(.pr$Series$to_r(self)), "in $to_vector():")
+Series_to_vector = \(int64_conversion = pl$options$int64_conversion ) {
+  unlist(unwrap(.pr$Series$to_r(self, int64_conversion )), "in $to_vector():")
 }
 
 #' Alias to Series_to_vector (backward compatibility)
@@ -298,11 +300,12 @@ Series_to_r_vector = Series_to_vector
 #' @rdname Series_to_r
 #' @name Series_to_r_list
 #' @description return R list (implicit as.list)
+#' @inheritParams pl_set_options
 #' @return R list
 #' @keywords Series
 #' @examples #
-Series_to_r_list = \() {
-  as.list(unwrap(.pr$Series$to_r(self)), "in $to_r_list():")
+Series_to_r_list = \(int64_conversion = pl$options$int64_conversion ) {
+  as.list(unwrap(.pr$Series$to_r(self, int64_conversion )), "in $to_r_list():")
 }