datafusion_functions/string/
overlay.rsuse std::any::Any;
use std::sync::Arc;
use arrow::array::{ArrayRef, GenericStringArray, OffsetSizeTrait};
use arrow::datatypes::DataType;
use crate::utils::{make_scalar_function, utf8_to_str_type};
use datafusion_common::cast::{
as_generic_string_array, as_int64_array, as_string_view_array,
};
use datafusion_common::{exec_err, Result};
use datafusion_expr::{ColumnarValue, Documentation, TypeSignature, Volatility};
use datafusion_expr::{ScalarUDFImpl, Signature};
use datafusion_macros::user_doc;
#[user_doc(
doc_section(label = "String Functions"),
description = "Returns the string which is replaced by another string from the specified position and specified count length.",
syntax_example = "overlay(str PLACING substr FROM pos [FOR count])",
sql_example = r#"```sql
> select overlay('Txxxxas' placing 'hom' from 2 for 4);
+--------------------------------------------------------+
| overlay(Utf8("Txxxxas"),Utf8("hom"),Int64(2),Int64(4)) |
+--------------------------------------------------------+
| Thomas |
+--------------------------------------------------------+
```"#,
standard_argument(name = "str", prefix = "String"),
argument(name = "substr", description = "Substring to replace in str."),
argument(
name = "pos",
description = "The start position to start the replace in str."
),
argument(
name = "count",
description = "The count of characters to be replaced from start position of str. If not specified, will use substr length instead."
)
)]
#[derive(Debug)]
pub struct OverlayFunc {
signature: Signature,
}
impl Default for OverlayFunc {
fn default() -> Self {
Self::new()
}
}
impl OverlayFunc {
pub fn new() -> Self {
use DataType::*;
Self {
signature: Signature::one_of(
vec![
TypeSignature::Exact(vec![Utf8View, Utf8View, Int64, Int64]),
TypeSignature::Exact(vec![Utf8, Utf8, Int64, Int64]),
TypeSignature::Exact(vec![LargeUtf8, LargeUtf8, Int64, Int64]),
TypeSignature::Exact(vec![Utf8View, Utf8View, Int64]),
TypeSignature::Exact(vec![Utf8, Utf8, Int64]),
TypeSignature::Exact(vec![LargeUtf8, LargeUtf8, Int64]),
],
Volatility::Immutable,
),
}
}
}
impl ScalarUDFImpl for OverlayFunc {
fn as_any(&self) -> &dyn Any {
self
}
fn name(&self) -> &str {
"overlay"
}
fn signature(&self) -> &Signature {
&self.signature
}
fn return_type(&self, arg_types: &[DataType]) -> Result<DataType> {
utf8_to_str_type(&arg_types[0], "overlay")
}
fn invoke_batch(
&self,
args: &[ColumnarValue],
_number_rows: usize,
) -> Result<ColumnarValue> {
match args[0].data_type() {
DataType::Utf8View | DataType::Utf8 => {
make_scalar_function(overlay::<i32>, vec![])(args)
}
DataType::LargeUtf8 => make_scalar_function(overlay::<i64>, vec![])(args),
other => exec_err!("Unsupported data type {other:?} for function overlay"),
}
}
fn documentation(&self) -> Option<&Documentation> {
self.doc()
}
}
macro_rules! process_overlay {
($string_array:expr, $characters_array:expr, $pos_num:expr) => {{
$string_array
.iter()
.zip($characters_array.iter())
.zip($pos_num.iter())
.map(|((string, characters), start_pos)| {
match (string, characters, start_pos) {
(Some(string), Some(characters), Some(start_pos)) => {
let string_len = string.chars().count();
let characters_len = characters.chars().count();
let replace_len = characters_len as i64;
let mut res =
String::with_capacity(string_len.max(characters_len));
if start_pos > 1 && start_pos - 1 < string_len as i64 {
let start = (start_pos - 1) as usize;
res.push_str(&string[..start]);
}
res.push_str(characters);
if start_pos + replace_len - 1 < string_len as i64 {
let end = (start_pos + replace_len - 1) as usize;
res.push_str(&string[end..]);
}
Ok(Some(res))
}
_ => Ok(None),
}
})
.collect::<Result<GenericStringArray<T>>>()
}};
($string_array:expr, $characters_array:expr, $pos_num:expr, $len_num:expr) => {{
$string_array
.iter()
.zip($characters_array.iter())
.zip($pos_num.iter())
.zip($len_num.iter())
.map(|(((string, characters), start_pos), len)| {
match (string, characters, start_pos, len) {
(Some(string), Some(characters), Some(start_pos), Some(len)) => {
let string_len = string.chars().count();
let characters_len = characters.chars().count();
let replace_len = len.min(string_len as i64);
let mut res =
String::with_capacity(string_len.max(characters_len));
if start_pos > 1 && start_pos - 1 < string_len as i64 {
let start = (start_pos - 1) as usize;
res.push_str(&string[..start]);
}
res.push_str(characters);
if start_pos + replace_len - 1 < string_len as i64 {
let end = (start_pos + replace_len - 1) as usize;
res.push_str(&string[end..]);
}
Ok(Some(res))
}
_ => Ok(None),
}
})
.collect::<Result<GenericStringArray<T>>>()
}};
}
fn overlay<T: OffsetSizeTrait>(args: &[ArrayRef]) -> Result<ArrayRef> {
let use_string_view = args[0].data_type() == &DataType::Utf8View;
if use_string_view {
string_view_overlay::<T>(args)
} else {
string_overlay::<T>(args)
}
}
pub fn string_overlay<T: OffsetSizeTrait>(args: &[ArrayRef]) -> Result<ArrayRef> {
match args.len() {
3 => {
let string_array = as_generic_string_array::<T>(&args[0])?;
let characters_array = as_generic_string_array::<T>(&args[1])?;
let pos_num = as_int64_array(&args[2])?;
let result = process_overlay!(string_array, characters_array, pos_num)?;
Ok(Arc::new(result) as ArrayRef)
}
4 => {
let string_array = as_generic_string_array::<T>(&args[0])?;
let characters_array = as_generic_string_array::<T>(&args[1])?;
let pos_num = as_int64_array(&args[2])?;
let len_num = as_int64_array(&args[3])?;
let result =
process_overlay!(string_array, characters_array, pos_num, len_num)?;
Ok(Arc::new(result) as ArrayRef)
}
other => {
exec_err!("overlay was called with {other} arguments. It requires 3 or 4.")
}
}
}
pub fn string_view_overlay<T: OffsetSizeTrait>(args: &[ArrayRef]) -> Result<ArrayRef> {
match args.len() {
3 => {
let string_array = as_string_view_array(&args[0])?;
let characters_array = as_string_view_array(&args[1])?;
let pos_num = as_int64_array(&args[2])?;
let result = process_overlay!(string_array, characters_array, pos_num)?;
Ok(Arc::new(result) as ArrayRef)
}
4 => {
let string_array = as_string_view_array(&args[0])?;
let characters_array = as_string_view_array(&args[1])?;
let pos_num = as_int64_array(&args[2])?;
let len_num = as_int64_array(&args[3])?;
let result =
process_overlay!(string_array, characters_array, pos_num, len_num)?;
Ok(Arc::new(result) as ArrayRef)
}
other => {
exec_err!("overlay was called with {other} arguments. It requires 3 or 4.")
}
}
}
#[cfg(test)]
mod tests {
use arrow::array::{Int64Array, StringArray};
use super::*;
#[test]
fn to_overlay() -> Result<()> {
let string =
Arc::new(StringArray::from(vec!["123", "abcdefg", "xyz", "Txxxxas"]));
let replace_string =
Arc::new(StringArray::from(vec!["abc", "qwertyasdfg", "ijk", "hom"]));
let start = Arc::new(Int64Array::from(vec![4, 1, 1, 2])); let end = Arc::new(Int64Array::from(vec![5, 7, 2, 4])); let res = overlay::<i32>(&[string, replace_string, start, end]).unwrap();
let result = as_generic_string_array::<i32>(&res).unwrap();
let expected = StringArray::from(vec!["abc", "qwertyasdfg", "ijkz", "Thomas"]);
assert_eq!(&expected, result);
Ok(())
}
}