romainfrancois commented on a change in pull request #7819:
URL: https://github.com/apache/arrow/pull/7819#discussion_r472991503
##########
File path: r/src/array_from_vector.cpp
##########
@@ -1064,42 +1062,42 @@ class FixedSizeBinaryVectorConverter : public
VectorConverter {
FixedSizeBinaryBuilder* typed_builder_;
};
-template <typename Builder>
+template <typename StringBuilder>
class StringVectorConverter : public VectorConverter {
public:
~StringVectorConverter() {}
Status Init(ArrayBuilder* builder) {
- typed_builder_ = checked_cast<Builder*>(builder);
+ typed_builder_ = checked_cast<StringBuilder*>(builder);
return Status::OK();
}
Status Ingest(SEXP obj) {
ARROW_RETURN_IF(TYPEOF(obj) != STRSXP,
Status::RError("Expecting a character vector"));
- R_xlen_t n = XLENGTH(obj);
- // Reserve enough space before appending
- int64_t size = 0;
- for (R_xlen_t i = 0; i < n; i++) {
- SEXP string_i = STRING_ELT(obj, i);
- if (string_i != NA_STRING) {
- size += XLENGTH(Rf_mkCharCE(Rf_translateCharUTF8(string_i), CE_UTF8));
- }
+ cpp11::strings s(obj);
+ RETURN_NOT_OK(typed_builder_->Reserve(s.size()));
+
+ // note: the total length is calculated without utf8
+ // conversion, so see this more as a hint rather than
+ // the actual total length
Review comment:
I've changed the strategy to use conversion upfront (only when needed)
so that we can get a real total size, and then use UnsafeAppend
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
