Struct GroupBy

pub struct GroupBy<'df> {
    pub df: &'df DataFrame,
    /* private fields */
}

Available on crate feature algorithm_group_by only.

Returned by a group_by operation on a DataFrame. This struct supports several aggregations.

Until described otherwise, the examples in this struct are performed on the following DataFrame:

use polars_core::prelude::*;

let dates = &[
"2020-08-21",
"2020-08-21",
"2020-08-22",
"2020-08-23",
"2020-08-22",
];
// date format
let fmt = "%Y-%m-%d";
// create date series
let s0 = DateChunked::parse_from_str_slice("date", dates, fmt)
        .into_series();
// create temperature series
let s1 = Series::new("temp".into(), [20, 10, 7, 9, 1]);
// create rain series
let s2 = Series::new("rain".into(), [0.2, 0.1, 0.3, 0.1, 0.01]);
// create a new DataFrame
let df = DataFrame::new(vec![s0, s1, s2]).unwrap();
println!("{:?}", df);

Outputs:

+------------+------+------+
| date       | temp | rain |
| ---        | ---  | ---  |
| Date       | i32  | f64  |
+============+======+======+
| 2020-08-21 | 20   | 0.2  |
+------------+------+------+
| 2020-08-21 | 10   | 0.1  |
+------------+------+------+
| 2020-08-22 | 7    | 0.3  |
+------------+------+------+
| 2020-08-23 | 9    | 0.1  |
+------------+------+------+
| 2020-08-22 | 1    | 0.01 |
+------------+------+------+

Fields

df: &'df DataFrame

Implementations

impl<'df> GroupBy<'df>

pub fn new( df: &'df DataFrame, by: Vec<Column>, groups: GroupsProxy, selected_agg: Option<Vec<PlSmallStr>>, ) -> Self

pub fn select<I: IntoIterator<Item = S>, S: Into<PlSmallStr>>( self, selection: I, ) -> Self

Select the column(s) that should be aggregated. You can select a single column or a slice of columns.

Note that making a selection with this method is not required. If you skip it all columns (except for the keys) will be selected for aggregation.

pub fn get_groups(&self) -> &GroupsProxy

Get the internal representation of the GroupBy operation. The Vec returned contains: (first_idx, Vec<indexes>) Where second value in the tuple is a vector with all matching indexes.

pub unsafe fn get_groups_mut(&mut self) -> &mut GroupsProxy

Get the internal representation of the GroupBy operation. The Vec returned contains: (first_idx, Vec<indexes>) Where second value in the tuple is a vector with all matching indexes.

Safety

Groups should always be in bounds of the DataFrame hold by this GroupBy. If you mutate it, you must hold that invariant.

pub fn take_groups(self) -> GroupsProxy

pub fn take_groups_mut(&mut self) -> GroupsProxy

pub fn keys_sliced(&self, slice: Option<(i64, usize)>) -> Vec<Column>

pub fn keys(&self) -> Vec<Column>

pub fn mean(&self) -> PolarsResult<DataFrame>

👎Deprecated since 0.24.1: use polars.lazy aggregations

Aggregate grouped series and compute the mean per group.

Example

fn example(df: DataFrame) -> PolarsResult<DataFrame> {
    df.group_by(["date"])?.select(["temp", "rain"]).mean()
}

Returns:

+------------+-----------+-----------+
| date       | temp_mean | rain_mean |
| ---        | ---       | ---       |
| Date       | f64       | f64       |
+============+===========+===========+
| 2020-08-23 | 9         | 0.1       |
+------------+-----------+-----------+
| 2020-08-22 | 4         | 0.155     |
+------------+-----------+-----------+
| 2020-08-21 | 15        | 0.15      |
+------------+-----------+-----------+

pub fn sum(&self) -> PolarsResult<DataFrame>

👎Deprecated since 0.24.1: use polars.lazy aggregations

Aggregate grouped series and compute the sum per group.

Example

fn example(df: DataFrame) -> PolarsResult<DataFrame> {
    df.group_by(["date"])?.select(["temp"]).sum()
}

Returns:

+------------+----------+
| date       | temp_sum |
| ---        | ---      |
| Date       | i32      |
+============+==========+
| 2020-08-23 | 9        |
+------------+----------+
| 2020-08-22 | 8        |
+------------+----------+
| 2020-08-21 | 30       |
+------------+----------+

pub fn min(&self) -> PolarsResult<DataFrame>

👎Deprecated since 0.24.1: use polars.lazy aggregations

Aggregate grouped series and compute the minimal value per group.

Example

fn example(df: DataFrame) -> PolarsResult<DataFrame> {
    df.group_by(["date"])?.select(["temp"]).min()
}

Returns:

+------------+----------+
| date       | temp_min |
| ---        | ---      |
| Date       | i32      |
+============+==========+
| 2020-08-23 | 9        |
+------------+----------+
| 2020-08-22 | 1        |
+------------+----------+
| 2020-08-21 | 10       |
+------------+----------+

pub fn max(&self) -> PolarsResult<DataFrame>

👎Deprecated since 0.24.1: use polars.lazy aggregations

Aggregate grouped series and compute the maximum value per group.

Example

fn example(df: DataFrame) -> PolarsResult<DataFrame> {
    df.group_by(["date"])?.select(["temp"]).max()
}

Returns:

+------------+----------+
| date       | temp_max |
| ---        | ---      |
| Date       | i32      |
+============+==========+
| 2020-08-23 | 9        |
+------------+----------+
| 2020-08-22 | 7        |
+------------+----------+
| 2020-08-21 | 20       |
+------------+----------+

pub fn first(&self) -> PolarsResult<DataFrame>

👎Deprecated since 0.24.1: use polars.lazy aggregations

Aggregate grouped Series and find the first value per group.

Example

fn example(df: DataFrame) -> PolarsResult<DataFrame> {
    df.group_by(["date"])?.select(["temp"]).first()
}

Returns:

+------------+------------+
| date       | temp_first |
| ---        | ---        |
| Date       | i32        |
+============+============+
| 2020-08-23 | 9          |
+------------+------------+
| 2020-08-22 | 7          |
+------------+------------+
| 2020-08-21 | 20         |
+------------+------------+

pub fn last(&self) -> PolarsResult<DataFrame>

👎Deprecated since 0.24.1: use polars.lazy aggregations

Aggregate grouped Series and return the last value per group.

Example

fn example(df: DataFrame) -> PolarsResult<DataFrame> {
    df.group_by(["date"])?.select(["temp"]).last()
}

Returns:

+------------+------------+
| date       | temp_last |
| ---        | ---        |
| Date       | i32        |
+============+============+
| 2020-08-23 | 9          |
+------------+------------+
| 2020-08-22 | 1          |
+------------+------------+
| 2020-08-21 | 10         |
+------------+------------+

pub fn n_unique(&self) -> PolarsResult<DataFrame>

👎Deprecated since 0.24.1: use polars.lazy aggregations

Aggregate grouped Series by counting the number of unique values.

Example

fn example(df: DataFrame) -> PolarsResult<DataFrame> {
    df.group_by(["date"])?.select(["temp"]).n_unique()
}

Returns:

+------------+---------------+
| date       | temp_n_unique |
| ---        | ---           |
| Date       | u32           |
+============+===============+
| 2020-08-23 | 1             |
+------------+---------------+
| 2020-08-22 | 2             |
+------------+---------------+
| 2020-08-21 | 2             |
+------------+---------------+

pub fn quantile( &self, quantile: f64, method: QuantileMethod, ) -> PolarsResult<DataFrame>

👎Deprecated since 0.24.1: use polars.lazy aggregations

Aggregate grouped Series and determine the quantile per group.

Example

fn example(df: DataFrame) -> PolarsResult<DataFrame> {
    df.group_by(["date"])?.select(["temp"]).quantile(0.2, QuantileMethod::default())
}

pub fn median(&self) -> PolarsResult<DataFrame>

👎Deprecated since 0.24.1: use polars.lazy aggregations

Aggregate grouped Series and determine the median per group.

Example

fn example(df: DataFrame) -> PolarsResult<DataFrame> {
    df.group_by(["date"])?.select(["temp"]).median()
}

pub fn var(&self, ddof: u8) -> PolarsResult<DataFrame>

👎Deprecated since 0.24.1: use polars.lazy aggregations

Aggregate grouped Series and determine the variance per group.

pub fn std(&self, ddof: u8) -> PolarsResult<DataFrame>

👎Deprecated since 0.24.1: use polars.lazy aggregations

Aggregate grouped Series and determine the standard deviation per group.

pub fn count(&self) -> PolarsResult<DataFrame>

Aggregate grouped series and compute the number of values per group.

Example

fn example(df: DataFrame) -> PolarsResult<DataFrame> {
    df.group_by(["date"])?.select(["temp"]).count()
}

Returns:

+------------+------------+
| date       | temp_count |
| ---        | ---        |
| Date       | u32        |
+============+============+
| 2020-08-23 | 1          |
+------------+------------+
| 2020-08-22 | 2          |
+------------+------------+
| 2020-08-21 | 2          |
+------------+------------+

pub fn groups(&self) -> PolarsResult<DataFrame>

Get the group_by group indexes.

Example

fn example(df: DataFrame) -> PolarsResult<DataFrame> {
    df.group_by(["date"])?.groups()
}

Returns:

+--------------+------------+
| date         | groups     |
| ---          | ---        |
| Date(days)   | list [u32] |
+==============+============+
| 2020-08-23   | "[3]"      |
+--------------+------------+
| 2020-08-22   | "[2, 4]"   |
+--------------+------------+
| 2020-08-21   | "[0, 1]"   |
+--------------+------------+

pub fn agg_list(&self) -> PolarsResult<DataFrame>

👎Deprecated since 0.24.1: use polars.lazy aggregations

Aggregate the groups of the group_by operation into lists.

Example

fn example(df: DataFrame) -> PolarsResult<DataFrame> {
    // GroupBy and aggregate to Lists
    df.group_by(["date"])?.select(["temp"]).agg_list()
}

Returns:

+------------+------------------------+
| date       | temp_agg_list          |
| ---        | ---                    |
| Date       | list [i32]             |
+============+========================+
| 2020-08-23 | "[Some(9)]"            |
+------------+------------------------+
| 2020-08-22 | "[Some(7), Some(1)]"   |
+------------+------------------------+
| 2020-08-21 | "[Some(20), Some(10)]" |
+------------+------------------------+

pub fn par_apply<F>(&self, f: F) -> PolarsResult<DataFrame>

where F: Fn(DataFrame) -> PolarsResult<DataFrame> + Send + Sync,

👎Deprecated since 0.24.1: use polars.lazy aggregations

Apply a closure over the groups as a new DataFrame in parallel.

pub fn apply<F>(&self, f: F) -> PolarsResult<DataFrame>

where F: FnMut(DataFrame) -> PolarsResult<DataFrame> + Send + Sync,

Apply a closure over the groups as a new DataFrame.

Trait Implementations

impl<'df> Clone for GroupBy<'df>

fn clone(&self) -> GroupBy<'df>

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<'df> Debug for GroupBy<'df>

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

Formats the value using the given formatter.