Struct NullState

pub struct NullState { /* private fields */ }

Track the accumulator null state per row: if any values for that group were null and if any values have been seen at all for that group.

This is part of the inner loop for many GroupsAccumulators, and thus the performance is critical and so there are multiple specialized implementations, invoked depending on the specific combinations of the input.

Typically there are 4 potential combinations of inputs must be special cased for performance:

• With / Without filter
• With / Without nulls in the input

If the input has nulls, then the accumulator must potentially handle each input null value specially (e.g. for SUM to mark the corresponding sum as null)

If there are filters present, NullState tracks if it has seen any value for that group (as some values may be filtered out). Without a filter, the accumulator is only passed groups that had at least one value to accumulate so they do not need to track if they have seen values for a particular group.

Implementations

impl NullState

pub fn new() -> Self

pub fn size(&self) -> usize

return the size of all buffers allocated by this null state, not including self

pub fn accumulate<T, F>( &mut self, group_indices: &[usize], values: &PrimitiveArray<T>, opt_filter: Option<&BooleanArray>, total_num_groups: usize, value_fn: F, )

where T: ArrowPrimitiveType + Send, F: FnMut(usize, T::Native) + Send,

Invokes value_fn(group_index, value) for each non null, non filtered value of value, while tracking which groups have seen null inputs and which groups have seen any inputs if necessary

Arguments:

• values: the input arguments to the accumulator
• group_indices: To which groups do the rows in values belong, (aka group_index)
• opt_filter: if present, only rows for which is Some(true) are included
• value_fn: function invoked for (group_index, value) where value is non null

See accumulate, for more details on how value_fn is called

When value_fn is called it also sets

1. self.seen_values[group_index] to true for all rows that had a non null value

pub fn accumulate_boolean<F>( &mut self, group_indices: &[usize], values: &BooleanArray, opt_filter: Option<&BooleanArray>, total_num_groups: usize, value_fn: F, )

where F: FnMut(usize, bool) + Send,

Invokes value_fn(group_index, value) for each non null, non filtered value in values, while tracking which groups have seen null inputs and which groups have seen any inputs, for BooleanArrays.

Since BooleanArray is not a PrimitiveArray it must be handled specially.

See Self::accumulate, which handles PrimitiveArrays, for more details on other arguments.

pub fn build(&mut self, emit_to: EmitTo) -> NullBuffer

Creates the a NullBuffer representing which group_indices should have null values (because they never saw any values) for the emit_to rows.

resets the internal state appropriately

Trait Implementations

impl Debug for NullState

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

Formats the value using the given formatter.

impl Default for NullState

fn default() -> Self

Returns the “default value” for a type.