arrow

GH-36367: [C++] Add a zipped range utility (#36393)



### Rationale for this change

We write a lot of loops over parallel ranges. This function is a proven pattern improving code clarity in other languages

### What changes are included in this PR?

A zip utility is added which simplifies writing loops over parallel ranges.

```diff
@@ -1118,9 +1118,8 @@ Status GetFieldsFromArray(const RjArray& json_fields, FieldPosition parent_pos,
                           DictionaryMemo* dictionary_memo,
                           std::vector<std::shared_ptr<Field>>* fields) {
   fields->resize(json_fields.Size());
-  for (rj::SizeType i = 0; i < json_fields.Size(); ++i) {
-    RETURN_NOT_OK(GetField(json_fields[i], parent_pos.child(static_cast<int>(i)),
-                           dictionary_memo, &(*fields)[i]));
+  for (auto [json_field, field, i] : Zip(json_fields, *fields, Enumerate<int>)) {
+    RETURN_NOT_OK(GetField(json_field, parent_pos.child(i), dictionary_memo, &field));
   }
   return Status::OK();
 }
```

Some(most) of the loops in json_internal.cc are rewritten to showcase usage.

### Are these changes tested?

Yes, in `range_test.cc`.

### Are there any user-facing changes?

No, this is an internal utility.

* Closes: #36367

Authored-by: Benjamin Kietzman <bengilgit@gmail.com>
Signed-off-by: Antoine Pitrou <antoine@python.org>