This is an automated email from the ASF dual-hosted git repository.
alamb 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 6ff8cc4d04 Support for `Arc<str>` in `ParquetRecordWriter` derive
macro (#8973)
6ff8cc4d04 is described below
commit 6ff8cc4d04750332208931b483c4703025ced079
Author: Ivan Reshetnikov <[email protected]>
AuthorDate: Thu Dec 11 20:47:37 2025 +0100
Support for `Arc<str>` in `ParquetRecordWriter` derive macro (#8973)
# Which issue does this PR close?
- Closes #8972.
# Rationale for this change
See #8972 for more info.
# What changes are included in this PR?
Updates the derive macro implementation to include string serialization
from `Arc<str>`.
# Are these changes tested?
Yes. I've updated `parquet_derive_test` to include writing `Arc<str>`.
And we've been using our fork with these changes in production for ~3
years.
# Are there any user-facing changes?
Not sure, probably not?
---------
Co-authored-by: Harry Bairstow <[email protected]>
---
parquet_derive/src/parquet_field.rs | 4 ++--
parquet_derive_test/src/lib.rs | 9 +++++++++
2 files changed, 11 insertions(+), 2 deletions(-)
diff --git a/parquet_derive/src/parquet_field.rs
b/parquet_derive/src/parquet_field.rs
index f51c22f65f..7473f23055 100644
--- a/parquet_derive/src/parquet_field.rs
+++ b/parquet_derive/src/parquet_field.rs
@@ -643,7 +643,7 @@ impl Type {
}
"f32" => BasicType::FLOAT,
"f64" => BasicType::DOUBLE,
- "String" | "str" => BasicType::BYTE_ARRAY,
+ "String" | "str" | "Arc < str >" => BasicType::BYTE_ARRAY,
"Uuid" => BasicType::FIXED_LEN_BYTE_ARRAY,
f => unimplemented!("{} currently is not supported", f),
}
@@ -733,7 +733,7 @@ impl Type {
"NaiveDate" => quote! { Some(LogicalType::Date) },
"NaiveDateTime" => quote! { None },
"f32" | "f64" => quote! { None },
- "String" | "str" => quote! { Some(LogicalType::String) },
+ "String" | "str" | "Arc < str >" => quote! {
Some(LogicalType::String) },
"Uuid" => quote! { Some(LogicalType::Uuid) },
f => unimplemented!("{} currently is not supported", f),
}
diff --git a/parquet_derive_test/src/lib.rs b/parquet_derive_test/src/lib.rs
index 71eb636069..fe96fa0e61 100644
--- a/parquet_derive_test/src/lib.rs
+++ b/parquet_derive_test/src/lib.rs
@@ -23,6 +23,7 @@
#![allow(clippy::approx_constant)]
use parquet_derive::{ParquetRecordReader, ParquetRecordWriter};
+use std::sync::Arc;
#[derive(ParquetRecordWriter)]
struct ACompleteRecord<'a> {
@@ -30,8 +31,10 @@ struct ACompleteRecord<'a> {
pub a_str: &'a str,
pub a_string: String,
pub a_borrowed_string: &'a String,
+ pub a_arc_str: Arc<str>,
pub maybe_a_str: Option<&'a str>,
pub maybe_a_string: Option<String>,
+ pub maybe_a_arc_str: Option<Arc<str>>,
pub i16: i16,
pub i32: i32,
pub u64: u64,
@@ -130,8 +133,10 @@ mod tests {
REQUIRED BINARY a_str (STRING);
REQUIRED BINARY a_string (STRING);
REQUIRED BINARY a_borrowed_string (STRING);
+ REQUIRED BINARY a_arc_str (STRING);
OPTIONAL BINARY maybe_a_str (STRING);
OPTIONAL BINARY maybe_a_string (STRING);
+ OPTIONAL BINARY maybe_a_arc_str (STRING);
REQUIRED INT32 i16 (INTEGER(16,true));
REQUIRED INT32 i32;
REQUIRED INT64 u64 (INTEGER(64,false));
@@ -159,8 +164,10 @@ mod tests {
let a_str = "hello mother".to_owned();
let a_borrowed_string = "cool news".to_owned();
+ let a_arc_str: Arc<str> = "hello arc".into();
let maybe_a_string = Some("it's true, I'm a string".to_owned());
let maybe_a_str = Some(&a_str[..]);
+ let maybe_a_arc_str = Some(a_arc_str.clone());
let borrowed_byte_vec = vec![0x68, 0x69, 0x70];
let borrowed_maybe_byte_vec = Some(vec![0x71, 0x72]);
let borrowed_maybe_borrowed_byte_vec = Some(&borrowed_byte_vec[..]);
@@ -170,8 +177,10 @@ mod tests {
a_str: &a_str[..],
a_string: "hello father".into(),
a_borrowed_string: &a_borrowed_string,
+ a_arc_str,
maybe_a_str: Some(&a_str[..]),
maybe_a_string: Some(a_str.clone()),
+ maybe_a_arc_str,
i16: -45,
i32: 456,
u64: 4563424,
