Skip to content

Optimize numeric-to-string coercion in StringArrayDecoder #7274

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 4 commits into from
Closed

Optimize numeric-to-string coercion in StringArrayDecoder #7274

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
6a9401a
JSON Reader Faster Coercion of Primitives to String
ndemir Mar 11, 2025
900ed2e
added comment about safety
ndemir Mar 12, 2025
ea3e7ae
call write_number when TapeElement::Number
ndemir Mar 12, 2025
917bd7a
Revert "call write_number when TapeElement::Number"
ndemir Mar 12, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 19 additions & 4 deletions arrow-json/src/reader/string_array.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,7 @@ use arrow_array::{Array, GenericStringArray, OffsetSizeTrait};
use arrow_data::ArrayData;
use arrow_schema::ArrowError;
use std::marker::PhantomData;
use std::io::Write;

use crate::reader::tape::{Tape, TapeElement};
use crate::reader::ArrayDecoder;
Expand All @@ -30,15 +31,25 @@ const FALSE: &str = "false";
pub struct StringArrayDecoder<O: OffsetSizeTrait> {
coerce_primitive: bool,
phantom: PhantomData<O>,
number_buffer: Vec<u8>,
}

impl<O: OffsetSizeTrait> StringArrayDecoder<O> {
pub fn new(coerce_primitive: bool) -> Self {
Self {
coerce_primitive,
phantom: Default::default(),
number_buffer: Vec::with_capacity(32),
}
}

fn write_number<T: std::fmt::Display>(&mut self, n: T) -> &str {
self.number_buffer.clear();
write!(&mut self.number_buffer, "{}", n).unwrap();
// SAFETY: We only write ASCII characters (digits, signs, decimal points,
// exponent symbols) into `number_buffer`, which are guaranteed valid UTF-8.
unsafe { std::str::from_utf8_unchecked(&self.number_buffer) }
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think you can avoid unsafe here by using

    number_buffer: String,

This is similar to how @zhuqi-lucas did it here:

        // Temporary buffer to avoid per-iteration allocation for numeric types
        let mut tmp_buf = String::new();
...

}
}

impl<O: OffsetSizeTrait> ArrayDecoder for StringArrayDecoder<O> {
Expand Down Expand Up @@ -103,20 +114,24 @@ impl<O: OffsetSizeTrait> ArrayDecoder for StringArrayDecoder<O> {
TapeElement::I64(high) if coerce_primitive => match tape.get(p + 1) {
TapeElement::I32(low) => {
let val = ((high as i64) << 32) | (low as u32) as i64;
builder.append_value(val.to_string());
let s = self.write_number(val);
builder.append_value(s);
}
_ => unreachable!(),
},
TapeElement::I32(n) if coerce_primitive => {
builder.append_value(n.to_string());
let s = self.write_number(n);
builder.append_value(s);
}
TapeElement::F32(n) if coerce_primitive => {
builder.append_value(n.to_string());
let s = self.write_number(n);
builder.append_value(s);
}
TapeElement::F64(high) if coerce_primitive => match tape.get(p + 1) {
TapeElement::F32(low) => {
let val = f64::from_bits(((high as u64) << 32) | low as u64);
builder.append_value(val.to_string());
let s = self.write_number(val);
builder.append_value(s);
}
_ => unreachable!(),
},
Expand Down