refactor value_counts (#2781)

polars/polars-core/src/chunked_array/ops/mod.rs

@@ -484,13 +484,6 @@ pub trait ChunkUnique<T> {
         ))
     }
 
-    /// Count the unique values.
-    fn value_counts(&self) -> Result<DataFrame> {
-        Err(PolarsError::InvalidOperation(
-            "is_duplicated is not implemented for this dtype".into(),
-        ))
-    }
-
     /// The most occurring value(s). Can return multiple Values
     #[cfg(feature = "mode")]
     #[cfg_attr(docsrs, doc(cfg(feature = "mode")))]

polars/polars-core/src/chunked_array/ops/unique/mod.rs

@@ -6,7 +6,6 @@ use crate::chunked_array::object::ObjectType;
 use crate::datatypes::PlHashSet;
 use crate::frame::groupby::{GroupsProxy, IntoGroupsProxy};
 use crate::prelude::*;
-use crate::utils::NoNull;
 use rayon::prelude::*;
 use std::hash::Hash;
 
@@ -167,22 +166,6 @@ macro_rules! arg_unique_ca {
     }};
 }
 
-macro_rules! impl_value_counts {
-    ($self:expr) => {{
-        let group_tuples = $self.group_tuples(true, false).into_idx();
-        let values =
-            unsafe { $self.take_unchecked(group_tuples.iter().map(|t| t.0 as usize).into()) };
-        let mut counts: NoNull<IdxCa> = group_tuples
-            .into_iter()
-            .map(|(_, groups)| groups.len() as IdxSize)
-            .collect();
-        counts.rename("counts");
-        let cols = vec![values.into_series(), counts.into_inner().into_series()];
-        let df = DataFrame::new_no_checks(cols);
-        df.sort(&["counts"], true)
-    }};
-}
-
 impl<T> ChunkUnique<T> for ChunkedArray<T>
 where
     T: PolarsIntegerType,
@@ -206,11 +189,6 @@ where
         is_unique_duplicated!(self, true)
     }
 
-    // TODO! implement on series. Not worth the compile times here.
-    fn value_counts(&self) -> Result<DataFrame> {
-        impl_value_counts!(self)
-    }
-
     fn n_unique(&self) -> Result<usize> {
         if self.null_count() > 0 {
             Ok(fill_set(self.into_iter().flatten()).len() + 1)
@@ -245,10 +223,6 @@ impl ChunkUnique<Utf8Type> for Utf8Chunked {
         is_unique_duplicated!(self, true)
     }
 
-    fn value_counts(&self) -> Result<DataFrame> {
-        impl_value_counts!(self)
-    }
-
     fn n_unique(&self) -> Result<usize> {
         if self.null_count() > 0 {
             Ok(fill_set(self.into_iter().flatten()).len() + 1)
@@ -396,9 +370,6 @@ impl ChunkUnique<Float32Type> for Float32Chunked {
     fn is_duplicated(&self) -> Result<BooleanChunked> {
         self.bit_repr_small().is_duplicated()
     }
-    fn value_counts(&self) -> Result<DataFrame> {
-        impl_value_counts!(self)
-    }
 }
 
 impl ChunkUnique<Float64Type> for Float64Chunked {
@@ -421,9 +392,6 @@ impl ChunkUnique<Float64Type> for Float64Chunked {
     fn is_duplicated(&self) -> Result<BooleanChunked> {
         self.bit_repr_large().is_duplicated()
     }
-    fn value_counts(&self) -> Result<DataFrame> {
-        impl_value_counts!(self)
-    }
 }
 
 #[cfg(feature = "is_first")]

polars/polars-core/src/frame/groupby/proxy.rs

@@ -217,6 +217,19 @@ impl GroupsProxy {
         }
     }
 
+    pub fn group_lengths(&self, name: &str) -> IdxCa {
+        let ca: NoNull<IdxCa> = match self {
+            GroupsProxy::Idx(groups) => groups
+                .iter()
+                .map(|(_, groups)| groups.len() as IdxSize)
+                .collect_trusted(),
+            GroupsProxy::Slice(groups) => groups.iter().map(|g| g[1]).collect_trusted(),
+        };
+        let mut ca = ca.into_inner();
+        ca.rename(name);
+        ca
+    }
+
     #[cfg(feature = "private")]
     pub fn par_iter(&self) -> GroupsProxyParIter {
         GroupsProxyParIter::new(self)

polars/polars-core/src/series/implementations/categorical.rs

@@ -303,10 +303,6 @@ impl SeriesTrait for SeriesWrap<CategoricalChunked> {
         self.0.cast(data_type)
     }
 
-    fn value_counts(&self) -> Result<DataFrame> {
-        self.0.value_counts()
-    }
-
     fn get(&self, index: usize) -> AnyValue {
         self.0.get_any_value(index)
     }

polars/polars-core/src/series/implementations/dates_time.rs

@@ -391,10 +391,6 @@ macro_rules! impl_dyn_series {
                 self.0.to_dummies()
             }
 
-            fn value_counts(&self) -> Result<DataFrame> {
-                self.0.value_counts()
-            }
-
             fn get(&self, index: usize) -> AnyValue {
                 self.0.get_any_value(index)
             }

polars/polars-core/src/series/implementations/datetime.rs

@@ -397,10 +397,6 @@ impl SeriesTrait for SeriesWrap<DatetimeChunked> {
         self.0.to_dummies()
     }
 
-    fn value_counts(&self) -> Result<DataFrame> {
-        self.0.value_counts()
-    }
-
     fn get(&self, index: usize) -> AnyValue {
         self.0.get_any_value(index)
     }

polars/polars-core/src/series/implementations/duration.rs

@@ -388,10 +388,6 @@ impl SeriesTrait for SeriesWrap<DurationChunked> {
         self.0.to_dummies()
     }
 
-    fn value_counts(&self) -> Result<DataFrame> {
-        self.0.value_counts()
-    }
-
     fn get(&self, index: usize) -> AnyValue {
         self.0.get_any_value(index)
     }

polars/polars-core/src/series/implementations/floats.rs

@@ -413,10 +413,6 @@ macro_rules! impl_dyn_series {
                 ToDummies::to_dummies(&self.0)
             }
 
-            fn value_counts(&self) -> Result<DataFrame> {
-                ChunkUnique::value_counts(&self.0)
-            }
-
             fn get(&self, index: usize) -> AnyValue {
                 self.0.get_any_value(index)
             }

polars/polars-core/src/series/implementations/mod.rs

@@ -580,10 +580,6 @@ macro_rules! impl_dyn_series {
                 ToDummies::to_dummies(&self.0)
             }
 
-            fn value_counts(&self) -> Result<DataFrame> {
-                ChunkUnique::value_counts(&self.0)
-            }
-
             fn get(&self, index: usize) -> AnyValue {
                 self.0.get_any_value(index)
             }

polars/polars-core/src/series/implementations/object.rs

@@ -170,10 +170,6 @@ where
         ))
     }
 
-    fn value_counts(&self) -> Result<DataFrame> {
-        ChunkUnique::value_counts(&self.0)
-    }
-
     fn get(&self, index: usize) -> AnyValue {
         ObjectChunked::get_any_value(&self.0, index)
     }

polars/polars-core/src/series/implementations/utf8.rs

@@ -252,10 +252,6 @@ impl SeriesTrait for SeriesWrap<Utf8Chunked> {
         ToDummies::to_dummies(&self.0)
     }
 
-    fn value_counts(&self) -> Result<DataFrame> {
-        ChunkUnique::value_counts(&self.0)
-    }
-
     fn get(&self, index: usize) -> AnyValue {
         self.0.get_any_value(index)
     }

polars/polars-core/src/series/ops/mod.rs

@@ -11,6 +11,7 @@ pub mod pct_change;
 #[cfg(feature = "round_series")]
 mod round;
 mod to_list;
+mod unique;
 
 #[derive(Copy, Clone)]
 pub enum NullBehavior {

polars/polars-core/src/series/ops/unique.rs

@@ -0,0 +1,14 @@
+use crate::prelude::*;
+
+impl Series {
+    /// Create a [`DataFrame`] with the unique `values` of this [`Series`] and a column `"counts"`
+    /// with dtype [`IdxType`]
+    pub fn value_counts(&self) -> Result<DataFrame> {
+        let groups = self.group_tuples(true, false);
+        let values = self.agg_first(&groups);
+        let counts = groups.group_lengths("counts");
+        let cols = vec![values.into_series(), counts.into_series()];
+        let df = DataFrame::new_no_checks(cols);
+        df.sort(&["counts"], true)
+    }
+}

polars/polars-core/src/series/series_trait.rs

@@ -607,10 +607,6 @@ pub trait SeriesTrait:
         invalid_operation_panic!(self)
     }
 
-    fn value_counts(&self) -> Result<DataFrame> {
-        invalid_operation_panic!(self)
-    }
-
     /// Get a single value by index. Don't use this operation for loops as a runtime cast is
     /// needed for every iteration.
     fn get(&self, _index: usize) -> AnyValue {