risingwave_sqlsmith/sql_gen/
types.rs

1// Copyright 2025 RisingWave 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
15//! This module contains datatypes and functions which can be generated by sqlsmith.
16
17use std::collections::{HashMap, HashSet};
18use std::sync::LazyLock;
19
20use itertools::Itertools;
21use risingwave_common::types::{DataType, DataTypeName};
22use risingwave_expr::aggregate::PbAggKind;
23use risingwave_expr::sig::{FUNCTION_REGISTRY, FuncSign};
24use risingwave_frontend::expr::{CastContext, CastSig as RwCastSig, ExprType, cast_sigs};
25use risingwave_sqlparser::ast::{BinaryOperator, DataType as AstDataType, StructField};
26
27pub(super) fn data_type_to_ast_data_type(data_type: &DataType) -> AstDataType {
28    match data_type {
29        DataType::Boolean => AstDataType::Boolean,
30        DataType::Int16 => AstDataType::SmallInt,
31        DataType::Int32 => AstDataType::Int,
32        DataType::Int64 => AstDataType::BigInt,
33        DataType::Int256 => AstDataType::Custom(vec!["rw_int256".into()].into()),
34        DataType::Serial => unreachable!("serial should not be generated"),
35        DataType::Decimal => AstDataType::Decimal(None, None),
36        DataType::Float32 => AstDataType::Real,
37        DataType::Float64 => AstDataType::Double,
38        DataType::Varchar => AstDataType::Varchar,
39        DataType::Bytea => AstDataType::Bytea,
40        DataType::Date => AstDataType::Date,
41        DataType::Timestamp => AstDataType::Timestamp(false),
42        DataType::Timestamptz => AstDataType::Timestamp(true),
43        DataType::Time => AstDataType::Time(false),
44        DataType::Interval => AstDataType::Interval,
45        DataType::Jsonb => AstDataType::Custom(vec!["JSONB".into()].into()),
46        DataType::Struct(inner) => AstDataType::Struct(
47            inner
48                .iter()
49                .map(|(name, typ)| StructField {
50                    name: name.into(),
51                    data_type: data_type_to_ast_data_type(typ),
52                })
53                .collect(),
54        ),
55        DataType::List(typ) => AstDataType::Array(Box::new(data_type_to_ast_data_type(typ))),
56        DataType::Map(_) => todo!(),
57    }
58}
59
60fn data_type_name_to_ast_data_type(data_type_name: &DataTypeName) -> Option<DataType> {
61    use DataTypeName as T;
62    match data_type_name {
63        T::Boolean => Some(DataType::Boolean),
64        T::Int16 => Some(DataType::Int16),
65        T::Int32 => Some(DataType::Int32),
66        T::Int64 => Some(DataType::Int64),
67        T::Decimal => Some(DataType::Decimal),
68        T::Float32 => Some(DataType::Float32),
69        T::Float64 => Some(DataType::Float64),
70        T::Varchar => Some(DataType::Varchar),
71        T::Date => Some(DataType::Date),
72        T::Timestamp => Some(DataType::Timestamp),
73        T::Timestamptz => Some(DataType::Timestamptz),
74        T::Time => Some(DataType::Time),
75        T::Interval => Some(DataType::Interval),
76        _ => None,
77    }
78}
79
80/// Provide internal `CastSig` which can be used for `struct` and `list`.
81#[derive(Clone)]
82pub struct CastSig {
83    pub from_type: DataType,
84    pub to_type: DataType,
85    pub context: CastContext,
86}
87
88impl TryFrom<RwCastSig> for CastSig {
89    type Error = String;
90
91    fn try_from(value: RwCastSig) -> Result<Self, Self::Error> {
92        if let Some(from_type) = data_type_name_to_ast_data_type(&value.from_type)
93            && let Some(to_type) = data_type_name_to_ast_data_type(&value.to_type)
94        {
95            Ok(CastSig {
96                from_type,
97                to_type,
98                context: value.context,
99            })
100        } else {
101            Err(format!("unsupported cast sig: {:?}", value))
102        }
103    }
104}
105
106/// Function ban list.
107/// These functions should be generated eventually, by adding expression constraints.
108/// If we naively generate arguments for these functions, it will affect sqlsmith
109/// effectiveness, e.g. cause it to crash.
110static FUNC_BAN_LIST: LazyLock<HashSet<ExprType>> = LazyLock::new(|| {
111    [
112        // FIXME: https://github.com/risingwavelabs/risingwave/issues/8003
113        ExprType::Repeat,
114        // The format argument needs to be handled specially. It is still generated in `gen_special_func`.
115        ExprType::Decode,
116        // ENABLE: https://github.com/risingwavelabs/risingwave/issues/16293
117        ExprType::Sqrt,
118        // ENABLE: https://github.com/risingwavelabs/risingwave/issues/16293
119        ExprType::Pow,
120    ]
121    .into_iter()
122    .collect()
123});
124
125/// Table which maps functions' return types to possible function signatures.
126// ENABLE: https://github.com/risingwavelabs/risingwave/issues/5826
127// TODO: Create a `SPECIAL_FUNC` table.
128// Otherwise when we dump the function table, we won't include those functions in
129// gen_special_func.
130pub(crate) static FUNC_TABLE: LazyLock<HashMap<DataType, Vec<&'static FuncSign>>> =
131    LazyLock::new(|| {
132        let mut funcs = HashMap::<DataType, Vec<&'static FuncSign>>::new();
133        FUNCTION_REGISTRY
134            .iter_scalars()
135            .filter(|func| {
136                func.inputs_type.iter().all(|t| {
137                    t.is_exact()
138                        && t.as_exact() != &DataType::Timestamptz
139                        && t.as_exact() != &DataType::Serial
140                }) && func.ret_type.is_exact()
141                    && !FUNC_BAN_LIST.contains(&func.name.as_scalar())
142                    && !func.deprecated // deprecated functions are not accepted by frontend
143            })
144            .for_each(|func| {
145                funcs
146                    .entry(func.ret_type.as_exact().clone())
147                    .or_default()
148                    .push(func)
149            });
150        funcs
151    });
152
153/// Set of invariant functions
154// ENABLE: https://github.com/risingwavelabs/risingwave/issues/5826
155pub(crate) static INVARIANT_FUNC_SET: LazyLock<HashSet<ExprType>> = LazyLock::new(|| {
156    FUNCTION_REGISTRY
157        .iter_scalars()
158        .map(|sig| sig.name.as_scalar())
159        .counts()
160        .into_iter()
161        .filter(|(_key, count)| *count == 1)
162        .map(|(key, _)| key)
163        .collect()
164});
165
166/// Table which maps aggregate functions' return types to possible function signatures.
167// ENABLE: https://github.com/risingwavelabs/risingwave/issues/5826
168pub(crate) static AGG_FUNC_TABLE: LazyLock<HashMap<DataType, Vec<&'static FuncSign>>> =
169    LazyLock::new(|| {
170        let mut funcs = HashMap::<DataType, Vec<&'static FuncSign>>::new();
171        FUNCTION_REGISTRY
172            .iter_aggregates()
173            .filter(|func| {
174                func.inputs_type
175                    .iter()
176                    .all(|t| t.is_exact() && t.as_exact() != &DataType::Timestamptz && t.as_exact() != &DataType::Serial)
177                    && func.ret_type.is_exact()
178                    // Ignored functions
179                    && ![
180                        PbAggKind::InternalLastSeenValue, // Use internally
181                        PbAggKind::Sum0, // Used internally
182                        PbAggKind::BitAnd,
183                        PbAggKind::BitOr,
184                        PbAggKind::BoolAnd,
185                        PbAggKind::BoolOr,
186                        PbAggKind::PercentileCont,
187                        PbAggKind::PercentileDisc,
188                        PbAggKind::Mode,
189                        PbAggKind::ApproxPercentile, // ENABLE: https://github.com/risingwavelabs/risingwave/issues/16293
190                        PbAggKind::JsonbObjectAgg, // ENABLE: https://github.com/risingwavelabs/risingwave/issues/16293
191                        PbAggKind::StddevSamp, // ENABLE: https://github.com/risingwavelabs/risingwave/issues/16293
192                        PbAggKind::VarSamp, // ENABLE: https://github.com/risingwavelabs/risingwave/issues/16293
193                    ]
194                    .contains(&func.name.as_aggregate())
195                    // Exclude 2 phase agg global sum.
196                    // Sum(Int64) -> Int64.
197                    // Otherwise it conflicts with normal aggregation:
198                    // Sum(Int64) -> Decimal.
199                    // And sqlsmith will generate expressions with wrong types.
200                    && if func.name.as_aggregate() == PbAggKind::Sum {
201                       !(func.inputs_type[0].as_exact() == &DataType::Int64 && func.ret_type.as_exact() == &DataType::Int64)
202                    } else {
203                       true
204                    }
205            })
206            .for_each(|func| {
207                funcs.entry(func.ret_type.as_exact().clone()).or_default().push(func)
208            });
209        funcs
210    });
211
212/// Build a cast map from return types to viable cast-signatures.
213/// NOTE: We avoid cast from varchar to other datatypes apart from itself.
214/// This is because arbitrary strings may not be able to cast,
215/// creating large number of invalid queries.
216pub(crate) static EXPLICIT_CAST_TABLE: LazyLock<HashMap<DataType, Vec<CastSig>>> =
217    LazyLock::new(|| {
218        let mut casts = HashMap::<DataType, Vec<CastSig>>::new();
219        cast_sigs()
220            .filter_map(|cast| cast.try_into().ok())
221            .filter(|cast: &CastSig| cast.context == CastContext::Explicit)
222            .filter(|cast| cast.from_type != DataType::Varchar || cast.to_type == DataType::Varchar)
223            .for_each(|cast| casts.entry(cast.to_type.clone()).or_default().push(cast));
224        casts
225    });
226
227/// Build a cast map from return types to viable cast-signatures.
228/// NOTE: We avoid cast from varchar to other datatypes apart from itself.
229/// This is because arbitrary strings may not be able to cast,
230/// creating large number of invalid queries.
231pub(crate) static IMPLICIT_CAST_TABLE: LazyLock<HashMap<DataType, Vec<CastSig>>> =
232    LazyLock::new(|| {
233        let mut casts = HashMap::<DataType, Vec<CastSig>>::new();
234        cast_sigs()
235            .filter_map(|cast| cast.try_into().ok())
236            .filter(|cast: &CastSig| cast.context == CastContext::Implicit)
237            .filter(|cast| cast.from_type != DataType::Varchar || cast.to_type == DataType::Varchar)
238            .for_each(|cast| casts.entry(cast.to_type.clone()).or_default().push(cast));
239        casts
240    });
241
242fn expr_type_to_inequality_op(typ: ExprType) -> Option<BinaryOperator> {
243    match typ {
244        ExprType::GreaterThan => Some(BinaryOperator::Gt),
245        ExprType::GreaterThanOrEqual => Some(BinaryOperator::GtEq),
246        ExprType::LessThan => Some(BinaryOperator::Lt),
247        ExprType::LessThanOrEqual => Some(BinaryOperator::LtEq),
248        ExprType::NotEqual => Some(BinaryOperator::NotEq),
249        _ => None,
250    }
251}
252
253/// Build set of binary inequality functions like `>`, `<`, etc...
254/// Maps from LHS, RHS argument to Inequality Operation
255/// For instance:
256/// GreaterThanOrEqual(Int16, Int64) -> Boolean
257/// Will store an entry of:
258/// Key: Int16, Int64
259/// Value: `BinaryOp::GreaterThanOrEqual`
260/// in the table.
261pub(crate) static BINARY_INEQUALITY_OP_TABLE: LazyLock<
262    HashMap<(DataType, DataType), Vec<BinaryOperator>>,
263> = LazyLock::new(|| {
264    let mut funcs = HashMap::<(DataType, DataType), Vec<BinaryOperator>>::new();
265    FUNCTION_REGISTRY
266        .iter_scalars()
267        .filter(|func| {
268            !FUNC_BAN_LIST.contains(&func.name.as_scalar())
269                && func.ret_type == DataType::Boolean.into()
270                && func.inputs_type.len() == 2
271                && func
272                    .inputs_type
273                    .iter()
274                    .all(|t| t.is_exact() && t.as_exact() != &DataType::Timestamptz)
275        })
276        .filter_map(|func| {
277            let lhs = func.inputs_type[0].as_exact().clone();
278            let rhs = func.inputs_type[1].as_exact().clone();
279            let op = expr_type_to_inequality_op(func.name.as_scalar())?;
280            Some(((lhs, rhs), op))
281        })
282        .for_each(|(args, op)| funcs.entry(args).or_default().push(op));
283    funcs
284});
risingwave_sqlsmith/sql_gen/types.rs

risingwave_sqlsmith/sql_gen/
types.rs
