Function get_index

datafusion/functions/src/regex/regexpinstr.rs:384–440

(
    value: Option<&str>,
    pattern: &'strings str,
    start: i64,
    n: i64,
    subexpr: i64,
    flags: Option<&'strings str>,
    regex_cache: &'cache mut HashMap<(&'strings str, Option<&'str

    }
}
fn get_index<'strings, 'cache>(
    value: Option<&str>,
    pattern: &'strings str,
    start: i64,
    n: i64,
    subexpr: i64,
    flags: Option<&'strings str>,
    regex_cache: &'cache mut HashMap<(&'strings str, Option<&'strings str>), Regex>,
) -> Result<Option<i64>, ArrowError>
where
    'strings: 'cache,
{
    let value = match value {
        None => return Ok(None),
        Some("") => return Ok(Some(0)),
        Some(value) => value,
    };
    let pattern: &Regex = compile_and_cache_regex(pattern, flags, regex_cache)?;
    // println!("get_index: value = {}, pattern = {}, start = {}, n = {}, subexpr = {}, flags = {:?}", value, pattern, start, n, subexpr, flags);
    if start < 1 {
        return Err(ArrowError::ComputeError(
            "regexp_instr() requires start to be 1-based".to_string(),
        ));
    }

    if n < 1 {
        return Err(ArrowError::ComputeError(
            "N must be 1 or greater".to_string(),
        ));
    }

    // --- Simplified byte_start_offset calculation ---
    let total_chars = value.chars().count() as i64;
    let byte_start_offset: usize = if start > total_chars {
        // If start is beyond the total characters, it means we start searching
        // after the string effectively. No matches possible.
        return Ok(Some(0));
    } else {
        // Get the byte offset for the (start - 1)-th character (0-based)
        value
            .char_indices()
            .nth((start - 1) as usize)
            .map(|(idx, _)| idx)
            .unwrap_or(0) // Should not happen if start is valid and <= total_chars
    };
    // --- End simplified calculation ---

    let search_slice = &value[byte_start_offset..];

    // Handle subexpression capturing first, as it takes precedence
    if subexpr > 0 {
        return handle_subexp(pattern, search_slice, subexpr, value, byte_start_offset);
    }

    // Use nth to get the N-th match (n is 1-based, nth is 0-based)
    get_nth_match(pattern, search_slice, n, byte_start_offset, value)
}