datafusion_functions/string/
bit_length.rsuse arrow::compute::kernels::length::bit_length;
use arrow::datatypes::DataType;
use std::any::Any;
use crate::utils::utf8_to_int_type;
use datafusion_common::{exec_err, Result, ScalarValue};
use datafusion_expr::{ColumnarValue, Documentation, Volatility};
use datafusion_expr::{ScalarUDFImpl, Signature};
use datafusion_macros::user_doc;
#[user_doc(
doc_section(label = "String Functions"),
description = "Returns the bit length of a string.",
syntax_example = "bit_length(str)",
sql_example = r#"```sql
> select bit_length('datafusion');
+--------------------------------+
| bit_length(Utf8("datafusion")) |
+--------------------------------+
| 80 |
+--------------------------------+
```"#,
standard_argument(name = "str", prefix = "String"),
related_udf(name = "length"),
related_udf(name = "octet_length")
)]
#[derive(Debug)]
pub struct BitLengthFunc {
signature: Signature,
}
impl Default for BitLengthFunc {
fn default() -> Self {
Self::new()
}
}
impl BitLengthFunc {
pub fn new() -> Self {
Self {
signature: Signature::string(1, Volatility::Immutable),
}
}
}
impl ScalarUDFImpl for BitLengthFunc {
fn as_any(&self) -> &dyn Any {
self
}
fn name(&self) -> &str {
"bit_length"
}
fn signature(&self) -> &Signature {
&self.signature
}
fn return_type(&self, arg_types: &[DataType]) -> Result<DataType> {
utf8_to_int_type(&arg_types[0], "bit_length")
}
fn invoke_batch(
&self,
args: &[ColumnarValue],
_number_rows: usize,
) -> Result<ColumnarValue> {
if args.len() != 1 {
return exec_err!(
"bit_length function requires 1 argument, got {}",
args.len()
);
}
match &args[0] {
ColumnarValue::Array(v) => Ok(ColumnarValue::Array(bit_length(v.as_ref())?)),
ColumnarValue::Scalar(v) => match v {
ScalarValue::Utf8(v) => Ok(ColumnarValue::Scalar(ScalarValue::Int32(
v.as_ref().map(|x| (x.len() * 8) as i32),
))),
ScalarValue::LargeUtf8(v) => Ok(ColumnarValue::Scalar(
ScalarValue::Int64(v.as_ref().map(|x| (x.len() * 8) as i64)),
)),
ScalarValue::Utf8View(v) => Ok(ColumnarValue::Scalar(
ScalarValue::Int32(v.as_ref().map(|x| (x.len() * 8) as i32)),
)),
_ => unreachable!("bit length"),
},
}
}
fn documentation(&self) -> Option<&Documentation> {
self.doc()
}
}