boolean.rs

use super::*;
use crate::chunked_array::comparison::*;
#[cfg(feature = "algorithm_group_by")]
use crate::frame::group_by::*;
use crate::prelude::*;

impl private::PrivateSeries for SeriesWrap<BooleanChunked> {
    fn compute_len(&mut self) {
        self.0.compute_len()
    }
    fn _field(&self) -> Cow<'_, Field> {
        Cow::Borrowed(self.0.ref_field())
    }
    fn _dtype(&self) -> &DataType {
        self.0.ref_field().dtype()
    }
    fn _get_flags(&self) -> StatisticsFlags {
        self.0.get_flags()
    }
    fn _set_flags(&mut self, flags: StatisticsFlags) {
        self.0.set_flags(flags)
    }

    unsafe fn equal_element(&self, idx_self: usize, idx_other: usize, other: &Series) -> bool {
        self.0.equal_element(idx_self, idx_other, other)
    }

    #[cfg(feature = "zip_with")]
    fn zip_with_same_type(&self, mask: &BooleanChunked, other: &Series) -> PolarsResult<Series> {
        ChunkZip::zip_with(&self.0, mask, other.as_ref().as_ref()).map(|ca| ca.into_series())
    }
    fn into_total_eq_inner<'a>(&'a self) -> Box<dyn TotalEqInner + 'a> {
        (&self.0).into_total_eq_inner()
    }
    fn into_total_ord_inner<'a>(&'a self) -> Box<dyn TotalOrdInner + 'a> {
        (&self.0).into_total_ord_inner()
    }

    fn vec_hash(
        &self,
        random_state: PlSeedableRandomStateQuality,
        buf: &mut Vec<u64>,
    ) -> PolarsResult<()> {
        self.0.vec_hash(random_state, buf)?;
        Ok(())
    }

    fn vec_hash_combine(
        &self,
        build_hasher: PlSeedableRandomStateQuality,
        hashes: &mut [u64],
    ) -> PolarsResult<()> {
        self.0.vec_hash_combine(build_hasher, hashes)?;
        Ok(())
    }

    #[cfg(feature = "algorithm_group_by")]
    unsafe fn agg_min(&self, groups: &GroupsType) -> Series {
        self.0.agg_min(groups)
    }

    #[cfg(feature = "algorithm_group_by")]
    unsafe fn agg_max(&self, groups: &GroupsType) -> Series {
        self.0.agg_max(groups)
    }

    #[cfg(feature = "algorithm_group_by")]
    unsafe fn agg_sum(&self, groups: &GroupsType) -> Series {
        self.0.agg_sum(groups)
    }

    #[cfg(feature = "algorithm_group_by")]
    unsafe fn agg_list(&self, groups: &GroupsType) -> Series {
        self.0.agg_list(groups)
    }
    #[cfg(feature = "algorithm_group_by")]
    unsafe fn agg_std(&self, groups: &GroupsType, _ddof: u8) -> Series {
        self.0
            .cast_with_options(&DataType::Float64, CastOptions::Overflowing)
            .unwrap()
            .agg_std(groups, _ddof)
    }
    #[cfg(feature = "algorithm_group_by")]
    unsafe fn agg_var(&self, groups: &GroupsType, _ddof: u8) -> Series {
        self.0
            .cast_with_options(&DataType::Float64, CastOptions::Overflowing)
            .unwrap()
            .agg_var(groups, _ddof)
    }

    #[cfg(feature = "bitwise")]
    unsafe fn agg_and(&self, groups: &GroupsType) -> Series {
        self.0.agg_and(groups).into_series()
    }
    #[cfg(feature = "bitwise")]
    unsafe fn agg_or(&self, groups: &GroupsType) -> Series {
        self.0.agg_or(groups).into_series()
    }
    #[cfg(feature = "bitwise")]
    unsafe fn agg_xor(&self, groups: &GroupsType) -> Series {
        self.0.agg_xor(groups).into_series()
    }

    #[cfg(feature = "algorithm_group_by")]
    fn group_tuples(&self, multithreaded: bool, sorted: bool) -> PolarsResult<GroupsType> {
        IntoGroupsType::group_tuples(&self.0, multithreaded, sorted)
    }

    fn arg_sort_multiple(
        &self,
        by: &[Column],
        options: &SortMultipleOptions,
    ) -> PolarsResult<IdxCa> {
        self.0.arg_sort_multiple(by, options)
    }
    fn add_to(&self, rhs: &Series) -> PolarsResult<Series> {
        NumOpsDispatch::add_to(&self.0, rhs)
    }
}

impl SeriesTrait for SeriesWrap<BooleanChunked> {
    fn rename(&mut self, name: PlSmallStr) {
        self.0.rename(name);
    }

    fn chunk_lengths(&self) -> ChunkLenIter<'_> {
        self.0.chunk_lengths()
    }
    fn name(&self) -> &PlSmallStr {
        self.0.name()
    }

    fn chunks(&self) -> &Vec<ArrayRef> {
        self.0.chunks()
    }
    unsafe fn chunks_mut(&mut self) -> &mut Vec<ArrayRef> {
        self.0.chunks_mut()
    }
    fn shrink_to_fit(&mut self) {
        self.0.shrink_to_fit()
    }

    fn slice(&self, offset: i64, length: usize) -> Series {
        self.0.slice(offset, length).into_series()
    }
    fn split_at(&self, offset: i64) -> (Series, Series) {
        let (a, b) = self.0.split_at(offset);
        (a.into_series(), b.into_series())
    }

    fn append(&mut self, other: &Series) -> PolarsResult<()> {
        polars_ensure!(self.0.dtype() == other.dtype(), append);
        self.0.append(other.as_ref().as_ref())?;
        Ok(())
    }
    fn append_owned(&mut self, other: Series) -> PolarsResult<()> {
        polars_ensure!(self.0.dtype() == other.dtype(), append);
        self.0.append_owned(other.take_inner())
    }

    fn extend(&mut self, other: &Series) -> PolarsResult<()> {
        polars_ensure!(self.0.dtype() == other.dtype(), extend);
        self.0.extend(other.as_ref().as_ref())?;
        Ok(())
    }

    fn filter(&self, filter: &BooleanChunked) -> PolarsResult<Series> {
        ChunkFilter::filter(&self.0, filter).map(|ca| ca.into_series())
    }

    fn _sum_as_f64(&self) -> f64 {
        self.0.sum().unwrap() as f64
    }

    fn mean(&self) -> Option<f64> {
        self.0.mean()
    }

    fn take(&self, indices: &IdxCa) -> PolarsResult<Series> {
        Ok(self.0.take(indices)?.into_series())
    }

    unsafe fn take_unchecked(&self, indices: &IdxCa) -> Series {
        self.0.take_unchecked(indices).into_series()
    }

    fn take_slice(&self, indices: &[IdxSize]) -> PolarsResult<Series> {
        Ok(self.0.take(indices)?.into_series())
    }

    unsafe fn take_slice_unchecked(&self, indices: &[IdxSize]) -> Series {
        self.0.take_unchecked(indices).into_series()
    }

    fn deposit(&self, validity: &Bitmap) -> Series {
        self.0.deposit(validity).into_series()
    }

    fn len(&self) -> usize {
        self.0.len()
    }

    fn rechunk(&self) -> Series {
        self.0.rechunk().into_owned().into_series()
    }

    fn new_from_index(&self, index: usize, length: usize) -> Series {
        ChunkExpandAtIndex::new_from_index(&self.0, index, length).into_series()
    }

    fn cast(&self, dtype: &DataType, options: CastOptions) -> PolarsResult<Series> {
        self.0.cast_with_options(dtype, options)
    }

    #[inline]
    unsafe fn get_unchecked(&self, index: usize) -> AnyValue<'_> {
        self.0.get_any_value_unchecked(index)
    }

    fn sort_with(&self, options: SortOptions) -> PolarsResult<Series> {
        Ok(ChunkSort::sort_with(&self.0, options).into_series())
    }

    fn arg_sort(&self, options: SortOptions) -> IdxCa {
        ChunkSort::arg_sort(&self.0, options)
    }

    fn null_count(&self) -> usize {
        self.0.null_count()
    }

    fn has_nulls(&self) -> bool {
        self.0.has_nulls()
    }

    #[cfg(feature = "algorithm_group_by")]
    fn unique(&self) -> PolarsResult<Series> {
        ChunkUnique::unique(&self.0).map(|ca| ca.into_series())
    }

    #[cfg(feature = "algorithm_group_by")]
    fn n_unique(&self) -> PolarsResult<usize> {
        ChunkUnique::n_unique(&self.0)
    }

    #[cfg(feature = "algorithm_group_by")]
    fn arg_unique(&self) -> PolarsResult<IdxCa> {
        ChunkUnique::arg_unique(&self.0)
    }

    fn unique_id(&self) -> PolarsResult<(IdxSize, Vec<IdxSize>)> {
        ChunkUnique::unique_id(&self.0)
    }

    fn is_null(&self) -> BooleanChunked {
        self.0.is_null()
    }

    fn is_not_null(&self) -> BooleanChunked {
        self.0.is_not_null()
    }

    fn reverse(&self) -> Series {
        ChunkReverse::reverse(&self.0).into_series()
    }

    fn as_single_ptr(&mut self) -> PolarsResult<usize> {
        self.0.as_single_ptr()
    }

    fn shift(&self, periods: i64) -> Series {
        ChunkShift::shift(&self.0, periods).into_series()
    }

    fn sum_reduce(&self) -> PolarsResult<Scalar> {
        Ok(ChunkAggSeries::sum_reduce(&self.0))
    }
    fn max_reduce(&self) -> PolarsResult<Scalar> {
        Ok(ChunkAggSeries::max_reduce(&self.0))
    }
    fn min_reduce(&self) -> PolarsResult<Scalar> {
        Ok(ChunkAggSeries::min_reduce(&self.0))
    }
    fn mean_reduce(&self) -> PolarsResult<Scalar> {
        Ok(Scalar::new(DataType::Float64, self.mean().into()))
    }
    fn median_reduce(&self) -> PolarsResult<Scalar> {
        let ca = self
            .0
            .cast_with_options(&DataType::Int8, CastOptions::Overflowing)
            .unwrap();
        let sc = ca.median_reduce()?;
        let v = sc.value().cast(&DataType::Float64);
        Ok(Scalar::new(DataType::Float64, v))
    }
    /// Get the variance of the Series as a new Series of length 1.
    fn var_reduce(&self, _ddof: u8) -> PolarsResult<Scalar> {
        let ca = self
            .0
            .cast_with_options(&DataType::Int8, CastOptions::Overflowing)
            .unwrap();
        let sc = ca.var_reduce(_ddof)?;
        let v = sc.value().cast(&DataType::Float64);
        Ok(Scalar::new(DataType::Float64, v))
    }
    /// Get the standard deviation of the Series as a new Series of length 1.
    fn std_reduce(&self, _ddof: u8) -> PolarsResult<Scalar> {
        let ca = self
            .0
            .cast_with_options(&DataType::Int8, CastOptions::Overflowing)
            .unwrap();
        let sc = ca.std_reduce(_ddof)?;
        let v = sc.value().cast(&DataType::Float64);
        Ok(Scalar::new(DataType::Float64, v))
    }
    fn and_reduce(&self) -> PolarsResult<Scalar> {
        let dt = DataType::Boolean;

        Ok(Scalar::new(
            dt,
            self.0
                .downcast_iter()
                .filter(|arr| !arr.is_empty())
                .filter_map(polars_compute::bitwise::BitwiseKernel::reduce_and)
                .reduce(|a, b| a & b)
                .map_or(AnyValue::Null, Into::into),
        ))
    }
    fn or_reduce(&self) -> PolarsResult<Scalar> {
        let dt = DataType::Boolean;

        Ok(Scalar::new(
            dt,
            self.0
                .downcast_iter()
                .filter(|arr| !arr.is_empty())
                .filter_map(polars_compute::bitwise::BitwiseKernel::reduce_or)
                .reduce(|a, b| a | b)
                .map_or(AnyValue::Null, Into::into),
        ))
    }
    fn xor_reduce(&self) -> PolarsResult<Scalar> {
        let dt = DataType::Boolean;

        Ok(Scalar::new(
            dt,
            self.0
                .downcast_iter()
                .filter(|arr| !arr.is_empty())
                .filter_map(polars_compute::bitwise::BitwiseKernel::reduce_xor)
                .reduce(|a, b| a ^ b)
                .map_or(AnyValue::Null, Into::into),
        ))
    }

    #[cfg(feature = "approx_unique")]
    fn approx_n_unique(&self) -> PolarsResult<IdxSize> {
        Ok(ChunkApproxNUnique::approx_n_unique(&self.0))
    }

    fn clone_inner(&self) -> Arc<dyn SeriesTrait> {
        Arc::new(SeriesWrap(Clone::clone(&self.0)))
    }

    fn find_validity_mismatch(&self, other: &Series, idxs: &mut Vec<IdxSize>) {
        self.0.find_validity_mismatch(other, idxs)
    }

    fn as_any(&self) -> &dyn Any {
        &self.0
    }

    fn as_any_mut(&mut self) -> &mut dyn Any {
        &mut self.0
    }

    fn as_phys_any(&self) -> &dyn Any {
        &self.0
    }

    fn as_arc_any(self: Arc<Self>) -> Arc<dyn Any + Send + Sync> {
        self as _
    }
}