polars-rows-iter 0.13.2

Library for easy and convenient row iteration of polars dataframes
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142

use crate::*;
use chrono::{DateTime, NaiveDateTime};
use polars::prelude::*;

impl<'a> IterFromColumn<'a> for NaiveDateTime {
    type RawInner = i64;
    fn create_iter(column: &'a Column) -> PolarsResult<impl Iterator<Item = Option<i64>> + 'a>
    where
        Self: Sized,
    {
        create_datetime_iter(column)
    }

    fn get_value(polars_value: Option<i64>, column_name: &str, dtype: &DataType) -> PolarsResult<Self>
    where
        Self: Sized,
    {
        let timestamp = polars_value
            .ok_or_else(|| <NaiveDateTime as IterFromColumn<'a>>::unexpected_null_value_error(column_name))?;

        create_datetime(timestamp, column_name, dtype)
    }
}

impl<'a> IterFromColumn<'a> for Option<NaiveDateTime> {
    type RawInner = i64;
    fn create_iter(column: &'a Column) -> PolarsResult<impl Iterator<Item = Option<i64>> + 'a>
    where
        Self: Sized,
    {
        create_datetime_iter(column)
    }

    fn get_value(polars_value: Option<i64>, column_name: &str, dtype: &DataType) -> PolarsResult<Self>
    where
        Self: Sized,
    {
        polars_value
            .map(|timestamp| create_datetime(timestamp, column_name, dtype))
            .transpose()
    }
}

fn create_datetime_iter<'a>(column: &'a Column) -> PolarsResult<impl Iterator<Item = Option<i64>> + 'a> {
    Ok(column.datetime()?.phys.iter())
}

fn create_datetime(timestamp: i64, column_name: &str, dtype: &DataType) -> PolarsResult<NaiveDateTime> {
    let (time_unit, _time_zone) = if let DataType::Datetime(tu, tz) = dtype {
        (tu, tz)
    } else {
        return Err(polars_err!(SchemaMismatch: "Unable to create chrono::DateTime from DataType: {dtype}"));
    };

    match time_unit {
        TimeUnit::Nanoseconds => Ok(DateTime::from_timestamp_nanos(timestamp).naive_utc()),
        TimeUnit::Microseconds => DateTime::from_timestamp_micros(timestamp).map(|dt|dt.naive_utc())
            .ok_or_else(|| polars_err!(OutOfBounds: "Value {timestamp} in column {column_name} is not a valid microseconds timestamp")),
        TimeUnit::Milliseconds => DateTime::from_timestamp_millis(timestamp).map(|dt|dt.naive_utc())
            .ok_or_else(|| polars_err!(OutOfBounds: "Value {timestamp} in column {column_name} is not a valid nanoseconds timestamp")),
    }
}

#[cfg(test)]
mod tests {

    use chrono::Utc;
    use itertools::Itertools;

    use super::*;

    #[test]
    fn datetime_should_be_read_from_datetime_millis() {
        let column_id = Column::new("id".into(), vec![1i32, 2, 3, 4, 5]);
        let column_datetime = Column::new(
            "dt".into(),
            vec![
                1732122821000i64,
                1700500421000,
                1668964421000,
                1637428421000,
                1605892421000,
            ],
        )
        .cast(&DataType::Datetime(TimeUnit::Milliseconds, None))
        .unwrap();
        let column_datetime_opt = Column::new(
            "dt_opt".into(),
            vec![
                Some(1605892421000i64),
                None,
                Some(1700500421000),
                None,
                Some(1637428421000),
            ],
        )
        .cast(&DataType::Datetime(TimeUnit::Milliseconds, None))
        .unwrap();

        let df = DataFrame::new(column_id.len(), vec![column_id, column_datetime, column_datetime_opt]).unwrap();

        #[derive(Debug, PartialEq, FromDataFrameRow)]
        struct TestRow {
            id: i32,
            dt: NaiveDateTime,
            dt_opt: Option<NaiveDateTime>,
        }

        let rows = df.rows_iter::<TestRow>().unwrap().map(|row| row.unwrap()).collect_vec();

        assert_eq!(
            rows,
            vec![
                TestRow {
                    id: 1,
                    dt: DateTime::<Utc>::from_timestamp(1732122821, 0).unwrap().naive_utc(),
                    dt_opt: Some(DateTime::<Utc>::from_timestamp(1605892421, 0).unwrap().naive_utc())
                },
                TestRow {
                    id: 2,
                    dt: DateTime::<Utc>::from_timestamp(1700500421, 0).unwrap().naive_utc(),
                    dt_opt: None
                },
                TestRow {
                    id: 3,
                    dt: DateTime::<Utc>::from_timestamp(1668964421, 0).unwrap().naive_utc(),
                    dt_opt: Some(DateTime::<Utc>::from_timestamp(1700500421, 0).unwrap().naive_utc())
                },
                TestRow {
                    id: 4,
                    dt: DateTime::<Utc>::from_timestamp(1637428421, 0).unwrap().naive_utc(),
                    dt_opt: None
                },
                TestRow {
                    id: 5,
                    dt: DateTime::<Utc>::from_timestamp(1605892421, 0).unwrap().naive_utc(),
                    dt_opt: Some(DateTime::<Utc>::from_timestamp(1637428421, 0).unwrap().naive_utc())
                }
            ]
        )
    }
}