This is an automated email from the ASF dual-hosted git repository.
Jefffrey pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/arrow-rs.git
The following commit(s) were added to refs/heads/main by this push:
new 3a35598892 fix: finish methods in `PrimitiveDictionaryBuilder` discard
type params (#11327)
3a35598892 is described below
commit 3a355988921a28b90855ea8a866223baf3f0e5e1
Author: Neil Conway <[email protected]>
AuthorDate: Thu Oct 1 22:12:55 2026 -0400
fix: finish methods in `PrimitiveDictionaryBuilder` discard type params
(#11327)
# Which issue does this PR close?
- Relates to #3972
# Rationale for this change
`PrimitiveDictionaryBuilder::finish_cloned` and `finish_preserve_values`
built the dictionary with `V::DATA_TYPE`, which is the default
`DataType` for the dictionary's value type. This loses type parameters,
like the scale/precision of a decimal value.
A similar bug was fixed for `finish` in #3972, but `finish_cloned` and
`finish_preserve_values` was missed.
# What changes are included in this PR?
See above.
# Are these changes tested?
Yes; existing tests pass, test expanded to cover this behavior. Verified
that the updated test fails without the behavior change.
# Are there any user-facing changes?
No.
# AI usage
Bug found and fix developed with Claude Code (Opus 5.5); I have reviewed
and understand the resulting code.
---
.../src/builder/primitive_dictionary_builder.rs | 21 ++++++++++++---------
1 file changed, 12 insertions(+), 9 deletions(-)
diff --git a/arrow-array/src/builder/primitive_dictionary_builder.rs
b/arrow-array/src/builder/primitive_dictionary_builder.rs
index bc5d878500..f2ab6a49d3 100644
--- a/arrow-array/src/builder/primitive_dictionary_builder.rs
+++ b/arrow-array/src/builder/primitive_dictionary_builder.rs
@@ -459,7 +459,8 @@ where
let values = self.values_builder.finish_cloned();
let keys = self.keys_builder.finish_cloned();
- let data_type = DataType::Dictionary(Box::new(K::DATA_TYPE),
Box::new(V::DATA_TYPE));
+ let data_type =
+ DataType::Dictionary(Box::new(K::DATA_TYPE),
Box::new(values.data_type().clone()));
let builder = keys
.into_data()
@@ -492,7 +493,8 @@ where
let values = self.values_builder.finish_cloned();
let keys = self.keys_builder.finish();
- let data_type = DataType::Dictionary(Box::new(K::DATA_TYPE),
Box::new(V::DATA_TYPE));
+ let data_type =
+ DataType::Dictionary(Box::new(K::DATA_TYPE),
Box::new(values.data_type().clone()));
let builder = keys
.into_data()
@@ -604,15 +606,16 @@ mod tests {
keys_builder,
values_builder,
);
+ let expected = DataType::Dictionary(
+ Box::new(DataType::Int32),
+ Box::new(DataType::Decimal128(1, 2)),
+ );
+ assert_eq!(builder.finish_cloned().data_type(), &expected);
+ assert_eq!(builder.finish_preserve_values().data_type(), &expected);
+
let dict_array = builder.finish();
assert_eq!(dict_array.value_type(), DataType::Decimal128(1, 2));
- assert_eq!(
- dict_array.data_type(),
- &DataType::Dictionary(
- Box::new(DataType::Int32),
- Box::new(DataType::Decimal128(1, 2)),
- )
- );
+ assert_eq!(dict_array.data_type(), &expected);
}
#[test]
