use crate::array::*;
use crate::compute::util::combine_option_bitmap;
use crate::error::{ArrowError, Result};
pub fn concat_elements_utf8<Offset: OffsetSizeTrait>(
left: &GenericStringArray<Offset>,
right: &GenericStringArray<Offset>,
) -> Result<GenericStringArray<Offset>> {
if left.len() != right.len() {
return Err(ArrowError::ComputeError(format!(
"Arrays must have the same length: {} != {}",
left.len(),
right.len()
)));
}
let output_bitmap = combine_option_bitmap(&[left.data(), right.data()], left.len())?;
let left_offsets = left.value_offsets();
let right_offsets = right.value_offsets();
let left_buffer = left.value_data();
let right_buffer = right.value_data();
let left_values = left_buffer.as_slice();
let right_values = right_buffer.as_slice();
let mut output_values = BufferBuilder::<u8>::new(
left_values.len() + right_values.len()
- left_offsets[0].to_usize().unwrap()
- right_offsets[0].to_usize().unwrap(),
);
let mut output_offsets = BufferBuilder::<Offset>::new(left_offsets.len());
output_offsets.append(Offset::zero());
for (left_idx, right_idx) in left_offsets.windows(2).zip(right_offsets.windows(2)) {
output_values.append_slice(
&left_values
[left_idx[0].to_usize().unwrap()..left_idx[1].to_usize().unwrap()],
);
output_values.append_slice(
&right_values
[right_idx[0].to_usize().unwrap()..right_idx[1].to_usize().unwrap()],
);
output_offsets.append(Offset::from_usize(output_values.len()).unwrap());
}
let builder = ArrayDataBuilder::new(GenericStringArray::<Offset>::get_data_type())
.len(left.len())
.add_buffer(output_offsets.finish())
.add_buffer(output_values.finish())
.null_bit_buffer(output_bitmap);
Ok(unsafe { builder.build_unchecked() }.into())
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_string_concat() {
let left = [Some("foo"), Some("bar"), None]
.into_iter()
.collect::<StringArray>();
let right = [None, Some("yyy"), Some("zzz")]
.into_iter()
.collect::<StringArray>();
let output = concat_elements_utf8(&left, &right).unwrap();
let expected = [None, Some("baryyy"), None]
.into_iter()
.collect::<StringArray>();
assert_eq!(output, expected);
}
#[test]
fn test_string_concat_empty_string() {
let left = [Some("foo"), Some(""), Some("bar")]
.into_iter()
.collect::<StringArray>();
let right = [Some("baz"), Some(""), Some("")]
.into_iter()
.collect::<StringArray>();
let output = concat_elements_utf8(&left, &right).unwrap();
let expected = [Some("foobaz"), Some(""), Some("bar")]
.into_iter()
.collect::<StringArray>();
assert_eq!(output, expected);
}
#[test]
fn test_string_concat_no_null() {
let left = StringArray::from(vec!["foo", "bar"]);
let right = StringArray::from(vec!["bar", "baz"]);
let output = concat_elements_utf8(&left, &right).unwrap();
let expected = StringArray::from(vec!["foobar", "barbaz"]);
assert_eq!(output, expected);
}
#[test]
fn test_string_concat_error() {
let left = StringArray::from(vec!["foo", "bar"]);
let right = StringArray::from(vec!["baz"]);
let output = concat_elements_utf8(&left, &right);
assert!(output.is_err());
}
#[test]
fn test_string_concat_slice() {
let left = &StringArray::from(vec![None, Some("foo"), Some("bar"), Some("baz")]);
let right = &StringArray::from(vec![Some("boo"), None, Some("far"), Some("faz")]);
let left_slice = left.slice(0, 3);
let right_slice = right.slice(1, 3);
let output = concat_elements_utf8(
left_slice
.as_any()
.downcast_ref::<GenericStringArray<i32>>()
.unwrap(),
right_slice
.as_any()
.downcast_ref::<GenericStringArray<i32>>()
.unwrap(),
)
.unwrap();
let expected = [None, Some("foofar"), Some("barfaz")]
.into_iter()
.collect::<StringArray>();
assert_eq!(output, expected);
let left_slice = left.slice(2, 2);
let right_slice = right.slice(1, 2);
let output = concat_elements_utf8(
left_slice
.as_any()
.downcast_ref::<GenericStringArray<i32>>()
.unwrap(),
right_slice
.as_any()
.downcast_ref::<GenericStringArray<i32>>()
.unwrap(),
)
.unwrap();
let expected = [None, Some("bazfar")].into_iter().collect::<StringArray>();
assert_eq!(output, expected);
}
}