databend_driver_core/
rows.rs

1// Copyright 2021 Datafuse Labs
2//
3// Licensed under the Apache License, Version 2.0 (the "License");
4// you may not use this file except in compliance with the License.
5// You may obtain a copy of the License at
6//
7//     http://www.apache.org/licenses/LICENSE-2.0
8//
9// Unless required by applicable law or agreed to in writing, software
10// distributed under the License is distributed on an "AS IS" BASIS,
11// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12// See the License for the specific language governing permissions and
13// limitations under the License.
14
15use std::pin::Pin;
16use std::task::Context;
17use std::task::Poll;
18
19use serde::Deserialize;
20use tokio_stream::{Stream, StreamExt};
21
22#[cfg(feature = "flight-sql")]
23use arrow::record_batch::RecordBatch;
24
25use crate::error::{Error, Result};
26use crate::schema::SchemaRef;
27use crate::value::Value;
28
29#[derive(Clone, Debug)]
30pub enum RowWithStats {
31    Row(Row),
32    Stats(ServerStats),
33}
34
35#[derive(Deserialize, Clone, Debug, Default)]
36pub struct ServerStats {
37    #[serde(default)]
38    pub total_rows: usize,
39    #[serde(default)]
40    pub total_bytes: usize,
41
42    #[serde(default)]
43    pub read_rows: usize,
44    #[serde(default)]
45    pub read_bytes: usize,
46
47    #[serde(default)]
48    pub write_rows: usize,
49    #[serde(default)]
50    pub write_bytes: usize,
51
52    #[serde(default)]
53    pub running_time_ms: f64,
54
55    #[serde(default)]
56    pub spill_file_nums: usize,
57
58    #[serde(default)]
59    pub spill_bytes: usize,
60}
61
62impl ServerStats {
63    pub fn normalize(&mut self) {
64        if self.total_rows == 0 {
65            self.total_rows = self.read_rows;
66        }
67        if self.total_bytes == 0 {
68            self.total_bytes = self.read_bytes;
69        }
70    }
71
72    pub fn merge(&mut self, other: &ServerStats) {
73        self.total_rows += other.total_rows;
74        self.total_bytes += other.total_bytes;
75        self.read_rows += other.read_rows;
76        self.read_bytes += other.read_bytes;
77        self.write_rows += other.write_rows;
78        self.write_bytes += other.write_bytes;
79        self.running_time_ms += other.running_time_ms;
80        self.spill_file_nums += other.spill_file_nums;
81        self.spill_bytes += other.spill_bytes;
82    }
83}
84
85impl From<databend_client::QueryStats> for ServerStats {
86    fn from(stats: databend_client::QueryStats) -> Self {
87        let mut p = Self {
88            total_rows: 0,
89            total_bytes: 0,
90            read_rows: stats.progresses.scan_progress.rows,
91            read_bytes: stats.progresses.scan_progress.bytes,
92            write_rows: stats.progresses.write_progress.rows,
93            write_bytes: stats.progresses.write_progress.bytes,
94            spill_file_nums: stats.progresses.spill_progress.file_nums,
95            spill_bytes: stats.progresses.spill_progress.bytes,
96            running_time_ms: stats.running_time_ms,
97        };
98        if let Some(total) = stats.progresses.total_scan {
99            p.total_rows = total.rows;
100            p.total_bytes = total.bytes;
101        }
102        p
103    }
104}
105
106#[derive(Clone, Debug, Default)]
107pub struct Row {
108    schema: SchemaRef,
109    values: Vec<Value>,
110}
111
112impl Row {
113    pub fn new(schema: SchemaRef, values: Vec<Value>) -> Self {
114        Self { schema, values }
115    }
116
117    pub fn len(&self) -> usize {
118        self.values.len()
119    }
120
121    pub fn is_empty(&self) -> bool {
122        self.values.is_empty()
123    }
124
125    pub fn values(&self) -> &[Value] {
126        &self.values
127    }
128
129    pub fn schema(&self) -> SchemaRef {
130        self.schema.clone()
131    }
132
133    pub fn from_vec(schema: SchemaRef, values: Vec<Value>) -> Self {
134        Self { schema, values }
135    }
136}
137
138impl TryFrom<(SchemaRef, Vec<Option<String>>)> for Row {
139    type Error = Error;
140
141    fn try_from((schema, data): (SchemaRef, Vec<Option<String>>)) -> Result<Self> {
142        let mut values: Vec<Value> = Vec::with_capacity(data.len());
143        for (field, val) in schema.fields().iter().zip(data.into_iter()) {
144            values.push(Value::try_from((&field.data_type, val))?);
145        }
146        Ok(Self::new(schema, values))
147    }
148}
149
150impl IntoIterator for Row {
151    type Item = Value;
152    type IntoIter = std::vec::IntoIter<Self::Item>;
153
154    fn into_iter(self) -> Self::IntoIter {
155        self.values.into_iter()
156    }
157}
158
159#[derive(Clone, Debug)]
160pub struct Rows {
161    rows: Vec<Row>,
162}
163
164impl Rows {
165    pub fn new(rows: Vec<Row>) -> Self {
166        Self { rows }
167    }
168
169    // pub fn schema(&self) -> SchemaRef {
170    //     self.schema.clone()
171    // }
172
173    pub fn rows(&self) -> &[Row] {
174        &self.rows
175    }
176
177    pub fn len(&self) -> usize {
178        self.rows.len()
179    }
180
181    pub fn is_empty(&self) -> bool {
182        self.rows.is_empty()
183    }
184}
185
186#[cfg(feature = "flight-sql")]
187impl TryFrom<RecordBatch> for Rows {
188    type Error = Error;
189    fn try_from(batch: RecordBatch) -> Result<Self> {
190        let batch_schema = batch.schema();
191        let schema = SchemaRef::new(batch_schema.clone().try_into()?);
192        let mut rows: Vec<Row> = Vec::new();
193        for i in 0..batch.num_rows() {
194            let mut values: Vec<Value> = Vec::new();
195            for j in 0..batch_schema.fields().len() {
196                let v = batch.column(j);
197                let field = batch_schema.field(j);
198                let value = Value::try_from((field, v, i))?;
199                values.push(value);
200            }
201            rows.push(Row::new(schema.clone(), values));
202        }
203        Ok(Self::new(rows))
204    }
205}
206
207impl IntoIterator for Rows {
208    type Item = Row;
209    type IntoIter = std::vec::IntoIter<Self::Item>;
210
211    fn into_iter(self) -> Self::IntoIter {
212        self.rows.into_iter()
213    }
214}
215
216macro_rules! replace_expr {
217    ($_t:tt $sub:expr) => {
218        $sub
219    };
220}
221
222// This macro implements TryFrom for tuple of types
223macro_rules! impl_tuple_from_row {
224    ( $($Ti:tt),+ ) => {
225        impl<$($Ti),+> TryFrom<Row> for ($($Ti,)+)
226        where
227            $($Ti: TryFrom<Value>),+
228        {
229            type Error = String;
230            fn try_from(row: Row) -> Result<Self, String> {
231                // It is not possible yet to get the number of metavariable repetitions
232                // ref: https://github.com/rust-lang/lang-team/issues/28#issue-644523674
233                // This is a workaround
234                let expected_len = <[()]>::len(&[$(replace_expr!(($Ti) ())),*]);
235
236                if expected_len != row.len() {
237                    return Err(format!("row size mismatch: expected {} columns, got {}", expected_len, row.len()));
238                }
239                let mut vals_iter = row.into_iter().enumerate();
240
241                Ok((
242                    $(
243                        {
244                            let (col_ix, col_value) = vals_iter
245                                .next()
246                                .unwrap(); // vals_iter size is checked before this code is reached,
247                                           // so it is safe to unwrap
248                            let t = col_value.get_type();
249                            $Ti::try_from(col_value)
250                                .map_err(|_| format!("failed converting column {} from type({:?}) to type({})", col_ix, t, std::any::type_name::<$Ti>()))?
251                        }
252                    ,)+
253                ))
254            }
255        }
256    }
257}
258
259// Implement FromRow for tuples of size up to 16
260impl_tuple_from_row!(T1);
261impl_tuple_from_row!(T1, T2);
262impl_tuple_from_row!(T1, T2, T3);
263impl_tuple_from_row!(T1, T2, T3, T4);
264impl_tuple_from_row!(T1, T2, T3, T4, T5);
265impl_tuple_from_row!(T1, T2, T3, T4, T5, T6);
266impl_tuple_from_row!(T1, T2, T3, T4, T5, T6, T7);
267impl_tuple_from_row!(T1, T2, T3, T4, T5, T6, T7, T8);
268impl_tuple_from_row!(T1, T2, T3, T4, T5, T6, T7, T8, T9);
269impl_tuple_from_row!(T1, T2, T3, T4, T5, T6, T7, T8, T9, T10);
270impl_tuple_from_row!(T1, T2, T3, T4, T5, T6, T7, T8, T9, T10, T11);
271impl_tuple_from_row!(T1, T2, T3, T4, T5, T6, T7, T8, T9, T10, T11, T12);
272impl_tuple_from_row!(T1, T2, T3, T4, T5, T6, T7, T8, T9, T10, T11, T12, T13);
273impl_tuple_from_row!(T1, T2, T3, T4, T5, T6, T7, T8, T9, T10, T11, T12, T13, T14);
274impl_tuple_from_row!(T1, T2, T3, T4, T5, T6, T7, T8, T9, T10, T11, T12, T13, T14, T15);
275impl_tuple_from_row!(T1, T2, T3, T4, T5, T6, T7, T8, T9, T10, T11, T12, T13, T14, T15, T16);
276
277pub struct RowIterator {
278    schema: SchemaRef,
279    it: Pin<Box<dyn Stream<Item = Result<Row>> + Send>>,
280}
281
282impl RowIterator {
283    pub fn new(schema: SchemaRef, it: Pin<Box<dyn Stream<Item = Result<Row>> + Send>>) -> Self {
284        Self { schema, it }
285    }
286
287    pub fn schema(&self) -> SchemaRef {
288        self.schema.clone()
289    }
290
291    pub async fn try_collect<T>(mut self) -> Result<Vec<T>>
292    where
293        T: TryFrom<Row>,
294        T::Error: std::fmt::Display,
295    {
296        let mut ret = Vec::new();
297        while let Some(row) = self.it.next().await {
298            let v = T::try_from(row?).map_err(|e| Error::Parsing(e.to_string()))?;
299            ret.push(v)
300        }
301        Ok(ret)
302    }
303}
304
305impl Stream for RowIterator {
306    type Item = Result<Row>;
307
308    fn poll_next(mut self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Option<Self::Item>> {
309        Pin::new(&mut self.it).poll_next(cx)
310    }
311}
312
313pub struct RowStatsIterator {
314    schema: SchemaRef,
315    it: Pin<Box<dyn Stream<Item = Result<RowWithStats>> + Send>>,
316}
317
318impl RowStatsIterator {
319    pub fn new(
320        schema: SchemaRef,
321        it: Pin<Box<dyn Stream<Item = Result<RowWithStats>> + Send>>,
322    ) -> Self {
323        Self { schema, it }
324    }
325
326    pub fn schema(&self) -> SchemaRef {
327        self.schema.clone()
328    }
329
330    pub async fn filter_rows(self) -> RowIterator {
331        let it = self.it.filter_map(|r| match r {
332            Ok(RowWithStats::Row(r)) => Some(Ok(r)),
333            Ok(_) => None,
334            Err(err) => Some(Err(err)),
335        });
336        RowIterator::new(self.schema, Box::pin(it))
337    }
338}
339
340impl Stream for RowStatsIterator {
341    type Item = Result<RowWithStats>;
342
343    fn poll_next(mut self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Option<Self::Item>> {
344        Pin::new(&mut self.it).poll_next(cx)
345    }
346}
databend_driver_core/rows.rs

databend_driver_core/
rows.rs
