Struct arrow::array::StringDictionaryBuilder

pub struct StringDictionaryBuilder<K> where
    K: ArrowDictionaryKeyType,  { /* fields omitted */ }

Array builder for DictionaryArray that stores Strings. For example to map a set of byte indices to String values. Note that the use of a HashMap here will not scale to very large arrays or result in an ordered dictionary.

use arrow::{
  array::{
    Int8Array, StringArray,
    PrimitiveBuilder, StringBuilder, StringDictionaryBuilder,
  },
  datatypes::Int8Type,
};

// Create a dictionary array indexed by bytes whose values are Strings.
// It can thus hold up to 256 distinct string values.

let key_builder = PrimitiveBuilder::<Int8Type>::new(100);
let value_builder = StringBuilder::new(100);
let mut builder = StringDictionaryBuilder::new(key_builder, value_builder);

// The builder builds the dictionary value by value
builder.append("abc").unwrap();
builder.append_null().unwrap();
builder.append("def").unwrap();
builder.append("def").unwrap();
builder.append("abc").unwrap();
let array = builder.finish();

assert_eq!(
  array.keys(),
  &Int8Array::from(vec![Some(0), None, Some(1), Some(1), Some(0)])
);

// Values are polymorphic and so require a downcast.
let av = array.values();
let ava: &StringArray = av.as_any().downcast_ref::<StringArray>().unwrap();

assert_eq!(ava.value(0), "abc");
assert_eq!(ava.value(1), "def");

Implementations

impl<K> StringDictionaryBuilder<K> where
    K: ArrowDictionaryKeyType, 

pub fn new(
    keys_builder: PrimitiveBuilder<K>,
    values_builder: StringBuilder
) -> Self

Creates a new StringDictionaryBuilder from a keys builder and a value builder.

pub fn new_with_dictionary(
    keys_builder: PrimitiveBuilder<K>,
    dictionary_values: &StringArray
) -> Result<Self>

Creates a new StringDictionaryBuilder from a keys builder and a dictionary which is initialized with the given values. The indices of those dictionary values are used as keys.

Example

use arrow::datatypes::Int16Type;
use arrow::array::{StringArray, StringDictionaryBuilder, PrimitiveBuilder, Int16Array};
use std::convert::TryFrom;

let dictionary_values = StringArray::from(vec![None, Some("abc"), Some("def")]);

let mut builder = StringDictionaryBuilder::new_with_dictionary(PrimitiveBuilder::<Int16Type>::new(3), &dictionary_values).unwrap();
builder.append("def").unwrap();
builder.append_null().unwrap();
builder.append("abc").unwrap();

let dictionary_array = builder.finish();

let keys = dictionary_array.keys();

assert_eq!(keys, &Int16Array::from(vec![Some(2), None, Some(1)]));

impl<K> StringDictionaryBuilder<K> where
    K: ArrowDictionaryKeyType, 

pub fn append(&mut self, value: impl AsRef<str>) -> Result<K::Native>

Append a primitive value to the array. Return an existing index if already present in the values array or a new index if the value is appended to the values array.

pub fn append_null(&mut self) -> Result<()>

pub fn finish(&mut self) -> DictionaryArray<K>

Builds the DictionaryArray and reset this builder.

Trait Implementations

impl<K> ArrayBuilder for StringDictionaryBuilder<K> where
    K: ArrowDictionaryKeyType, 

fn as_any(&self) -> &dyn Any

Returns the builder as an non-mutable Any reference.

fn as_any_mut(&mut self) -> &mut dyn Any

Returns the builder as an mutable Any reference.

fn into_box_any(self: Box<Self>) -> Box<dyn Any>

Returns the boxed builder as a box of Any.

fn len(&self) -> usize

Returns the number of array slots in the builder

fn is_empty(&self) -> bool

Returns whether the number of array slots is zero

fn finish(&mut self) -> ArrayRef

Builds the array and reset this builder.

impl<K: Debug> Debug for StringDictionaryBuilder<K> where
    K: ArrowDictionaryKeyType,
    K::Native: Debug, 

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.