Function arrow::compute::kernels::substring::substring_by_char

source · 
pub fn substring_by_char<OffsetSize>(
    array: &GenericByteArray<GenericStringType<OffsetSize>>,
    start: i64,
    length: Option<u64>,
) -> Result<GenericByteArray<GenericStringType<OffsetSize>>, ArrowError>
where
    OffsetSize: OffsetSizeTrait,
Expand description

§Arguments

  • array - The input string array

  • start - The start index of all substrings. If start >= 0, then count from the start of the string, otherwise count from the end of the string.

  • length(option) - The length of all substrings. If length is None, then the substring is from start to the end of the string.

Attention: Both start and length are counted by char.

§Performance

This function is slower than substring. Theoretically, the time complexity is O(n) where n is the length of the value buffer. It is recommended to use substring if the input array only contains ASCII chars.

§Basic usage

let array = StringArray::from(vec![Some("arrow"), None, Some("Γ ⊢x:T")]);
let result = substring_by_char(&array, 1, Some(4)).unwrap();
assert_eq!(result, StringArray::from(vec![Some("rrow"), None, Some(" ⊢x:")]));