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
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170

pub mod csv;
pub mod ipc;
pub mod json;
use crate::prelude::*;
use arrow::{
    csv::Reader as ArrowCsvReader, error::Result as ArrowResult, json::Reader as ArrowJsonReader,
    record_batch::RecordBatch,
};
use std::io::{Read, Seek, Write};
use std::sync::Arc;

pub trait SerReader<R>
where
    R: Read + Seek,
{
    fn new(reader: R) -> Self;

    /// Rechunk to a single chunk after Reading file.
    fn set_rechunk(self, rechunk: bool) -> Self;

    /// Take the SerReader and return a parsed DataFrame.
    fn finish(self) -> Result<DataFrame>;
}

pub trait SerWriter<'a, W>
where
    W: Write,
{
    fn new(writer: &'a mut W) -> Self;
    fn finish(self, df: &mut DataFrame) -> Result<()>;
}

pub trait ArrowReader {
    fn next(&mut self) -> ArrowResult<Option<RecordBatch>>;

    fn schema(&self) -> Arc<Schema>;
}

impl<R: Read> ArrowReader for ArrowCsvReader<R> {
    fn next(&mut self) -> ArrowResult<Option<RecordBatch>> {
        self.next()
    }

    fn schema(&self) -> Arc<Schema> {
        self.schema()
    }
}

impl<R: Read> ArrowReader for ArrowJsonReader<R> {
    fn next(&mut self) -> ArrowResult<Option<RecordBatch>> {
        self.next()
    }

    fn schema(&self) -> Arc<Schema> {
        self.schema()
    }
}

pub fn finish_reader<R: ArrowReader>(mut reader: R, rechunk: bool) -> Result<DataFrame> {
    let mut columns = reader
        .schema()
        .fields()
        .iter()
        .map(|field| match field.data_type() {
            ArrowDataType::UInt8 => {
                Series::UInt8(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::UInt16 => {
                Series::UInt16(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::UInt32 => {
                Series::UInt32(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::UInt64 => {
                Series::UInt64(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Int8 => {
                Series::Int8(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Int16 => {
                Series::Int16(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Int32 => {
                Series::Int32(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Int64 => {
                Series::Int64(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Float32 => {
                Series::Float32(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Float64 => {
                Series::Float64(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Utf8 => {
                Series::Utf8(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Boolean => {
                Series::Bool(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Date32(DateUnit::Millisecond) => {
                Series::Date32(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Date64(DateUnit::Millisecond) => {
                Series::Date64(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Duration(TimeUnit::Nanosecond) => {
                Series::DurationNanosecond(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Duration(TimeUnit::Microsecond) => {
                Series::DurationMicrosecond(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Duration(TimeUnit::Millisecond) => {
                Series::DurationMillisecond(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Duration(TimeUnit::Second) => {
                Series::DurationSecond(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Time64(TimeUnit::Nanosecond) => {
                Series::Time64Nanosecond(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Time64(TimeUnit::Microsecond) => {
                Series::Time64Microsecond(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Time32(TimeUnit::Millisecond) => {
                Series::Time32Millisecond(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Time32(TimeUnit::Second) => {
                Series::Time32Second(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Timestamp(TimeUnit::Nanosecond, _) => {
                Series::TimestampNanosecond(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Timestamp(TimeUnit::Microsecond, _) => {
                Series::TimestampMicrosecond(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Timestamp(TimeUnit::Millisecond, _) => {
                Series::TimestampMillisecond(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            ArrowDataType::Timestamp(TimeUnit::Second, _) => {
                Series::TimestampSecond(ChunkedArray::new_from_chunks(field.name(), vec![]))
            }
            _ => unimplemented!(),
        })
        .collect::<Vec<_>>();

    while let Some(batch) = reader.next()? {
        batch
            .columns()
            .into_iter()
            .zip(&mut columns)
            .map(|(arr, ser)| ser.append_array(arr.clone()))
            .collect::<Result<Vec<_>>>()?;
    }

    if rechunk {
        columns = columns
            .into_iter()
            .map(|s| {
                let s = s.rechunk(None)?;
                Ok(s)
            })
            .collect::<Result<Vec<_>>>()?;
    }

    Ok(DataFrame {
        schema: reader.schema(),
        columns,
    })
}