From 4c0a3f127cb921a246a88d752ae17b4d968cbd93 Mon Sep 17 00:00:00 2001
From: Kyle Husmann <kyle.husmann@gmail.com>
Date: Wed, 17 Jul 2024 15:23:28 -0700
Subject: [PATCH] add support for collector-level na args to the backend

---
 src/collectors.h | 7 +++++--
 src/columns.h    | 5 ++++-
 2 files changed, 9 insertions(+), 3 deletions(-)

diff --git a/src/collectors.h b/src/collectors.h
index 30fdad68..0854401c 100644
--- a/src/collectors.h
+++ b/src/collectors.h
@@ -54,6 +54,7 @@ class collector {
         type_(derive_type(cpp11::strings(data_.attr("class"))[0])),
         altrep_(altrep) {}
   column_type type() const { return type_; }
+  SEXP na() const { return data_["na"]; }
   SEXP name() const { return name_; }
   SEXP operator[](const char* nme) { return data_[nme]; }
   bool use_altrep() {
@@ -170,6 +171,8 @@ inline collectors resolve_collectors(
     std::string my_col_type = cpp11::strings(my_collector.attr("class"))[0];
 
     if (my_col_type == "collector_guess") {
+      auto my_col_na = my_collector["na"];
+      auto my_col_na_res = Rf_isNull(my_col_na) ? na : my_col_na;
       cpp11::writable::strings col_vals(guess_num);
       for (R_xlen_t j = 0; j < guess_num - 1; ++j) {
         size_t row = j * guess_step;
@@ -187,10 +190,10 @@ inline collectors resolve_collectors(
             locale_info->encoder_.makeSEXP(str.begin(), str.end(), true);
       }
 
-      auto type = guess_type__(col_vals, na, locale_info.get(), false);
+      auto type = guess_type__(col_vals, my_col_na_res, locale_info.get(), false);
       auto fun_name = std::string("col_") + type;
       auto col_type = vroom[fun_name.c_str()];
-      my_collectors[col] = col_type();
+      my_collectors[col] = col_type("na"_nm = my_col_na);
     }
   }
 
diff --git a/src/columns.h b/src/columns.h
index 68216e99..aee2037d 100644
--- a/src/columns.h
+++ b/src/columns.h
@@ -129,11 +129,14 @@ inline cpp11::list create_columns(
       continue;
     }
 
+    auto col_na = collector.na();
+    auto col_na_res = Rf_isNull(col_na) ? na : cpp11::strings(col_na);
+
     // This is deleted in the finalizers when the vectors are GC'd by R
     auto info = new vroom_vec_info{
         idx->get_column(col),
         num_threads,
-        std::make_shared<cpp11::strings>(na),
+        std::make_shared<cpp11::strings>(col_na_res),
         locale_info,
         *errors,
         std::string()};