From f5139f086ed2facf1fc6fc9f1b5a2410d19bc6e5 Mon Sep 17 00:00:00 2001
From: alpaylan <alpkeles99@gmail.com>
Date: Mon, 27 Jan 2025 01:24:20 +0300
Subject: [PATCH 1/5] add select-limit property

---
 simulator/generation/plan.rs     |  3 ++
 simulator/generation/property.rs | 64 ++++++++++++++++++++++++++++++++
 simulator/generation/query.rs    |  2 +
 simulator/model/query.rs         | 10 ++++-
 simulator/shrink/plan.rs         | 12 +++---
 5 files changed, 85 insertions(+), 6 deletions(-)

diff --git a/simulator/generation/plan.rs b/simulator/generation/plan.rs
index 2a794fbbc..2c8c940cf 100644
--- a/simulator/generation/plan.rs
+++ b/simulator/generation/plan.rs
@@ -282,6 +282,9 @@ impl Interactions {
                             query.shadow(env);
                         }
                     }
+                    Property::SelectLimit { select } => {
+                        select.shadow(env);
+                    },
                 }
                 for interaction in property.interactions() {
                     match interaction {
diff --git a/simulator/generation/property.rs b/simulator/generation/property.rs
index bfa1e1ed5..517baf372 100644
--- a/simulator/generation/property.rs
+++ b/simulator/generation/property.rs
@@ -62,6 +62,17 @@ pub(crate) enum Property {
         /// Additional interactions in the middle of the property
         queries: Vec<Query>,
     },
+    /// Select Limit is a property in which the select query
+    /// has a limit clause that is respected by the query.
+    /// The execution of the property is as follows
+    ///    SELECT * FROM <t> WHERE <predicate> LIMIT <n>
+    /// This property is a single-interaction property.
+    /// The interaction has the following constraints;
+    /// - The select query will respect the limit clause.
+    SelectLimit {
+        /// The select query
+        select: Select,
+    },
 }
 
 impl Property {
@@ -69,6 +80,7 @@ impl Property {
         match self {
             Property::InsertSelect { .. } => "Insert-Select".to_string(),
             Property::DoubleCreateFailure { .. } => "Double-Create-Failure".to_string(),
+            Property::SelectLimit { .. } => "Select-Limit".to_string(),
         }
     }
     /// interactions construct a list of interactions, which is an executable representation of the property.
@@ -164,6 +176,38 @@ impl Property {
 
                 interactions
             }
+            Property::SelectLimit { select } => {
+                let table_name = select.table.clone();
+
+                let assumption = Interaction::Assumption(Assertion {
+                    message: format!("table {} exists", table_name),
+                    func: Box::new({
+                        let table_name = table_name.clone();
+                        move |_: &Vec<ResultSet>, env: &SimulatorEnv| {
+                            Ok(env.tables.iter().any(|t| t.name == table_name))
+                        }
+                    }),
+                });
+
+                let limit = select.limit.clone().unwrap_or(0);
+
+                let assertion = Interaction::Assertion(Assertion {
+                    message: "select query should respect the limit clause".to_string(),
+                    func: Box::new(move |stack: &Vec<ResultSet>, _: &SimulatorEnv| {
+                        let last = stack.last().unwrap();
+                        match last {
+                            Ok(rows) => Ok(limit >= rows.len()),
+                            Err(_) => Ok(true),
+                        }
+                    }),
+                });
+
+                vec![
+                    assumption,
+                    Interaction::Query(Query::Select(select.clone())),
+                    assertion,
+                ]
+            }
         }
     }
 }
@@ -248,6 +292,7 @@ fn property_insert_select<R: rand::Rng>(
     let select_query = Select {
         table: table.name.clone(),
         predicate: Predicate::arbitrary_from(rng, (table, &row)),
+        limit: None,
     };
 
     Property::InsertSelect {
@@ -258,6 +303,21 @@ fn property_insert_select<R: rand::Rng>(
     }
 }
 
+fn property_select_limit<R: rand::Rng>(
+    rng: &mut R,
+    env: &SimulatorEnv,
+) -> Property {
+    // Get a random table
+    let table = pick(&env.tables, rng);
+    // Select the table
+    let select = Select {
+        table: table.name.clone(),
+        predicate: Predicate::arbitrary_from(rng, table),
+        limit: Some(rng.gen_range(1..=5)),
+    };
+    Property::SelectLimit { select }
+}
+
 fn property_double_create_failure<R: rand::Rng>(
     rng: &mut R,
     env: &SimulatorEnv,
@@ -312,6 +372,10 @@ impl ArbitraryFrom<(&SimulatorEnv, &InteractionStats)> for Property {
                     remaining_.create / 2.0,
                     Box::new(|rng: &mut R| property_double_create_failure(rng, env, &remaining_)),
                 ),
+                (
+                    remaining_.read,
+                    Box::new(|rng: &mut R| property_select_limit(rng, env)),
+                ),
             ],
             rng,
         )
diff --git a/simulator/generation/query.rs b/simulator/generation/query.rs
index 8b93fa993..1a0d167fc 100644
--- a/simulator/generation/query.rs
+++ b/simulator/generation/query.rs
@@ -23,6 +23,7 @@ impl ArbitraryFrom<&Vec<Table>> for Select {
         Self {
             table: table.name.clone(),
             predicate: Predicate::arbitrary_from(rng, table),
+            limit: Some(rng.gen_range(0..=1000)),
         }
     }
 }
@@ -33,6 +34,7 @@ impl ArbitraryFrom<&Vec<&Table>> for Select {
         Self {
             table: table.name.clone(),
             predicate: Predicate::arbitrary_from(rng, *table),
+            limit: Some(rng.gen_range(0..=1000)),
         }
     }
 }
diff --git a/simulator/model/query.rs b/simulator/model/query.rs
index f03bbde6f..8c861edc4 100644
--- a/simulator/model/query.rs
+++ b/simulator/model/query.rs
@@ -140,6 +140,7 @@ impl Create {
 pub(crate) struct Select {
     pub(crate) table: String,
     pub(crate) predicate: Predicate,
+    pub(crate) limit: Option<usize>,
 }
 
 impl Select {
@@ -190,7 +191,14 @@ impl Display for Query {
             Self::Select(Select {
                 table,
                 predicate: guard,
-            }) => write!(f, "SELECT * FROM {} WHERE {}", table, guard),
+                limit,
+            }) => write!(
+                f,
+                "SELECT * FROM {} WHERE {}{}",
+                table,
+                guard,
+                limit.map_or("".to_string(), |l| format!(" LIMIT {}", l))
+            ),
             Self::Insert(Insert { table, values }) => {
                 write!(f, "INSERT INTO {} VALUES ", table)?;
                 for (i, row) in values.iter().enumerate() {
diff --git a/simulator/shrink/plan.rs b/simulator/shrink/plan.rs
index 92867d82e..bc5eaa4d0 100644
--- a/simulator/shrink/plan.rs
+++ b/simulator/shrink/plan.rs
@@ -1,5 +1,8 @@
 use crate::{
-    generation::plan::{InteractionPlan, Interactions},
+    generation::{
+        plan::{InteractionPlan, Interactions},
+        property::Property,
+    },
     model::query::Query,
     runner::execution::Execution,
 };
@@ -27,12 +30,11 @@ impl InteractionPlan {
         for interaction in plan.plan.iter_mut() {
             if let Interactions::Property(p) = interaction {
                 match p {
-                    crate::generation::property::Property::InsertSelect { queries, .. }
-                    | crate::generation::property::Property::DoubleCreateFailure {
-                        queries, ..
-                    } => {
+                    Property::InsertSelect { queries, .. }
+                    | Property::DoubleCreateFailure { queries, .. } => {
                         queries.clear();
                     }
+                    Property::SelectLimit { .. } => {}
                 }
             }
         }

From 48d091e112757abaf091376fd84f06aca7a2dc37 Mon Sep 17 00:00:00 2001
From: alpaylan <alpkeles99@gmail.com>
Date: Sat, 1 Feb 2025 10:40:07 -0500
Subject: [PATCH 2/5] add insert into <table> <select> generation

---
 simulator/generation/mod.rs      |  35 +++++++-
 simulator/generation/plan.rs     |  32 ++++----
 simulator/generation/property.rs |  47 ++++++-----
 simulator/generation/query.rs    | 129 ++++++++++++++++--------------
 simulator/model/query.rs         | 133 ++++++++++++++++++++++---------
 simulator/shrink/plan.rs         |   2 +-
 6 files changed, 242 insertions(+), 136 deletions(-)

diff --git a/simulator/generation/mod.rs b/simulator/generation/mod.rs
index 04b5ee03d..71b8488be 100644
--- a/simulator/generation/mod.rs
+++ b/simulator/generation/mod.rs
@@ -1,4 +1,7 @@
-use std::{iter::Sum, ops::SubAssign};
+use std::{
+    iter::Sum,
+    ops::SubAssign,
+};
 
 use anarchist_readable_name_generator_lib::readable_name_custom;
 use rand::{distributions::uniform::SampleUniform, Rng};
@@ -60,6 +63,36 @@ pub(crate) fn one_of<'a, T, R: Rng>(choices: Vec<Box<dyn Fn(&mut R) -> T + 'a>>,
     choices[index](rng)
 }
 
+/// backtrack is a helper function for composing different "failable" generators.
+/// The function takes a list of functions that return an Option<T>, along with number of retries
+/// to make before giving up.
+pub(crate) fn backtrack<'a, T, R: Rng>(
+    mut choices: Vec<(u32, Box<dyn Fn(&mut R) -> Option<T> + 'a>)>,
+    rng: &mut R,
+) -> T {
+    loop {
+        // If there are no more choices left, we give up
+        let choices_ = choices
+            .iter()
+            .enumerate()
+            .filter(|(_, (retries, _))| *retries > 0)
+            .collect::<Vec<_>>();
+        if choices_.is_empty() {
+            panic!("backtrack: no more choices left");
+        }
+        // Run a one_of on the remaining choices
+        let (choice_index, choice) = pick(&choices_, rng);
+        let choice_index = *choice_index;
+        // If the choice returns None, we decrement the number of retries and try again
+        let result = choice.1(rng);
+        if let Some(result) = result {
+            return result;
+        } else {
+            choices[choice_index].0 -= 1;
+        }
+    }
+}
+
 /// pick is a helper function for uniformly picking a random element from a slice
 pub(crate) fn pick<'a, T, R: Rng>(choices: &'a [T], rng: &mut R) -> &'a T {
     let index = rng.gen_range(0..choices.len());
diff --git a/simulator/generation/plan.rs b/simulator/generation/plan.rs
index 2c8c940cf..a2651369d 100644
--- a/simulator/generation/plan.rs
+++ b/simulator/generation/plan.rs
@@ -14,10 +14,7 @@ use crate::{
 
 use crate::generation::{frequency, Arbitrary, ArbitraryFrom};
 
-use super::{
-    pick,
-    property::{remaining, Property},
-};
+use super::property::{remaining, Property};
 
 pub(crate) type ResultSet = Result<Vec<Vec<Value>>>;
 
@@ -261,7 +258,7 @@ impl Interactions {
         match self {
             Interactions::Property(property) => {
                 match property {
-                    Property::InsertSelect {
+                    Property::InsertValuesSelect {
                         insert,
                         row_index: _,
                         queries,
@@ -284,7 +281,7 @@ impl Interactions {
                     }
                     Property::SelectLimit { select } => {
                         select.shadow(env);
-                    },
+                    }
                 }
                 for interaction in property.interactions() {
                     match interaction {
@@ -295,12 +292,16 @@ impl Interactions {
                                 }
                             }
                             Query::Insert(insert) => {
+                                let values = match &insert {
+                                    Insert::Values { values, .. } => values.clone(),
+                                    Insert::Select { select, .. } => select.shadow(env),
+                                };
                                 let table = env
                                     .tables
                                     .iter_mut()
-                                    .find(|t| t.name == insert.table)
+                                    .find(|t| t.name == insert.table())
                                     .unwrap();
-                                table.rows.extend(insert.values.clone());
+                                table.rows.extend(values);
                             }
                             Query::Delete(_) => todo!(),
                             Query::Select(_) => {}
@@ -311,7 +312,9 @@ impl Interactions {
                     }
                 }
             }
-            Interactions::Query(query) => query.shadow(env),
+            Interactions::Query(query) => {
+                query.shadow(env);
+            }
             Interactions::Fault(_) => {}
         }
     }
@@ -392,12 +395,10 @@ impl ArbitraryFrom<&mut SimulatorEnv> for InteractionPlan {
 }
 
 impl Interaction {
-    pub(crate) fn shadow(&self, env: &mut SimulatorEnv) {
+    pub(crate) fn shadow(&self, env: &mut SimulatorEnv) -> Vec<Vec<Value>> {
         match self {
             Self::Query(query) => query.shadow(env),
-            Self::Assumption(_) => {}
-            Self::Assertion(_) => {}
-            Self::Fault(_) => {}
+            Self::Assumption(_) | Self::Assertion(_) | Self::Fault(_) => vec![],
         }
     }
     pub(crate) fn execute_query(&self, conn: &mut Rc<Connection>) -> ResultSet {
@@ -548,12 +549,11 @@ fn create_table<R: rand::Rng>(rng: &mut R, _env: &SimulatorEnv) -> Interactions
 }
 
 fn random_read<R: rand::Rng>(rng: &mut R, env: &SimulatorEnv) -> Interactions {
-    Interactions::Query(Query::Select(Select::arbitrary_from(rng, &env.tables)))
+    Interactions::Query(Query::Select(Select::arbitrary_from(rng, env)))
 }
 
 fn random_write<R: rand::Rng>(rng: &mut R, env: &SimulatorEnv) -> Interactions {
-    let table = pick(&env.tables, rng);
-    let insert_query = Query::Insert(Insert::arbitrary_from(rng, table));
+    let insert_query = Query::Insert(Insert::arbitrary_from(rng, env));
     Interactions::Query(insert_query)
 }
 
diff --git a/simulator/generation/property.rs b/simulator/generation/property.rs
index 517baf372..37c1fe744 100644
--- a/simulator/generation/property.rs
+++ b/simulator/generation/property.rs
@@ -3,7 +3,7 @@ use serde::{Deserialize, Serialize};
 
 use crate::{
     model::{
-        query::{Create, Delete, Insert, Predicate, Query, Select},
+        query::{Create, Delete, Distinctness, Insert, Predicate, Query, Select},
         table::Value,
     },
     runner::env::SimulatorEnv,
@@ -34,7 +34,7 @@ pub(crate) enum Property {
     /// - The inserted row will not be deleted.
     /// - The inserted row will not be updated.
     /// - The table `t` will not be renamed, dropped, or altered.
-    InsertSelect {
+    InsertValuesSelect {
         /// The insert query
         insert: Insert,
         /// Selected row index
@@ -78,7 +78,7 @@ pub(crate) enum Property {
 impl Property {
     pub(crate) fn name(&self) -> String {
         match self {
-            Property::InsertSelect { .. } => "Insert-Select".to_string(),
+            Property::InsertValuesSelect { .. } => "Insert-Values-Select".to_string(),
             Property::DoubleCreateFailure { .. } => "Double-Create-Failure".to_string(),
             Property::SelectLimit { .. } => "Select-Limit".to_string(),
         }
@@ -88,26 +88,33 @@ impl Property {
     /// and `interaction` cannot be serialized directly.
     pub(crate) fn interactions(&self) -> Vec<Interaction> {
         match self {
-            Property::InsertSelect {
+            Property::InsertValuesSelect {
                 insert,
                 row_index,
                 queries,
                 select,
             } => {
+                let (table, values) = if let Insert::Values { table, values } = insert {
+                    (table, values)
+                } else {
+                    unreachable!(
+                        "insert query should be Insert::Values for Insert-Values-Select property"
+                    )
+                };
                 // Check that the insert query has at least 1 value
                 assert!(
-                    !insert.values.is_empty(),
+                    !values.is_empty(),
                     "insert query should have at least 1 value"
                 );
 
                 // Pick a random row within the insert values
-                let row = insert.values[*row_index].clone();
+                let row = values[*row_index].clone();
 
                 // Assume that the table exists
                 let assumption = Interaction::Assumption(Assertion {
-                    message: format!("table {} exists", insert.table),
+                    message: format!("table {} exists", insert.table()),
                     func: Box::new({
-                        let table_name = insert.table.clone();
+                        let table_name = table.clone();
                         move |_: &Vec<ResultSet>, env: &SimulatorEnv| {
                             Ok(env.tables.iter().any(|t| t.name == table_name))
                         }
@@ -118,7 +125,7 @@ impl Property {
                     message: format!(
                         "row [{:?}] not found in table {}",
                         row.iter().map(|v| v.to_string()).collect::<Vec<String>>(),
-                        insert.table,
+                        insert.table(),
                     ),
                     func: Box::new(move |stack: &Vec<ResultSet>, _: &SimulatorEnv| {
                         let rows = stack.last().unwrap();
@@ -189,7 +196,7 @@ impl Property {
                     }),
                 });
 
-                let limit = select.limit.clone().unwrap_or(0);
+                let limit = select.limit.unwrap_or(0);
 
                 let assertion = Interaction::Assertion(Assertion {
                     message: "select query should respect the limit clause".to_string(),
@@ -212,6 +219,7 @@ impl Property {
     }
 }
 
+#[derive(Debug)]
 pub(crate) struct Remaining {
     pub(crate) read: f64,
     pub(crate) write: f64,
@@ -236,7 +244,7 @@ pub(crate) fn remaining(env: &SimulatorEnv, stats: &InteractionStats) -> Remaini
     }
 }
 
-fn property_insert_select<R: rand::Rng>(
+fn property_insert_values_select<R: rand::Rng>(
     rng: &mut R,
     env: &SimulatorEnv,
     remaining: &Remaining,
@@ -253,7 +261,7 @@ fn property_insert_select<R: rand::Rng>(
     let row = rows[row_index].clone();
 
     // Insert the rows
-    let insert_query = Insert {
+    let insert_query = Insert::Values {
         table: table.name.clone(),
         values: rows,
     };
@@ -265,7 +273,7 @@ fn property_insert_select<R: rand::Rng>(
     // - [ ] The inserted row will not be updated. (todo: add this constraint once UPDATE is implemented)
     // - [ ] The table `t` will not be renamed, dropped, or altered. (todo: add this constraint once ALTER or DROP is implemented)
     for _ in 0..rng.gen_range(0..3) {
-        let query = Query::arbitrary_from(rng, (table, remaining));
+        let query = Query::arbitrary_from(rng, (env, remaining));
         match &query {
             Query::Delete(Delete {
                 table: t,
@@ -293,9 +301,10 @@ fn property_insert_select<R: rand::Rng>(
         table: table.name.clone(),
         predicate: Predicate::arbitrary_from(rng, (table, &row)),
         limit: None,
+        distinct: Distinctness::All,
     };
 
-    Property::InsertSelect {
+    Property::InsertValuesSelect {
         insert: insert_query,
         row_index,
         queries,
@@ -303,10 +312,7 @@ fn property_insert_select<R: rand::Rng>(
     }
 }
 
-fn property_select_limit<R: rand::Rng>(
-    rng: &mut R,
-    env: &SimulatorEnv,
-) -> Property {
+fn property_select_limit<R: rand::Rng>(rng: &mut R, env: &SimulatorEnv) -> Property {
     // Get a random table
     let table = pick(&env.tables, rng);
     // Select the table
@@ -314,6 +320,7 @@ fn property_select_limit<R: rand::Rng>(
         table: table.name.clone(),
         predicate: Predicate::arbitrary_from(rng, table),
         limit: Some(rng.gen_range(1..=5)),
+        distinct: Distinctness::All,
     };
     Property::SelectLimit { select }
 }
@@ -336,7 +343,7 @@ fn property_double_create_failure<R: rand::Rng>(
     // - [x] There will be no errors in the middle interactions.(best effort)
     // - [ ] Table `t` will not be renamed or dropped.(todo: add this constraint once ALTER or DROP is implemented)
     for _ in 0..rng.gen_range(0..3) {
-        let query = Query::arbitrary_from(rng, (table, remaining));
+        let query = Query::arbitrary_from(rng, (env, remaining));
         match &query {
             Query::Create(Create { table: t }) => {
                 // There will be no errors in the middle interactions.
@@ -366,7 +373,7 @@ impl ArbitraryFrom<(&SimulatorEnv, &InteractionStats)> for Property {
             vec![
                 (
                     f64::min(remaining_.read, remaining_.write),
-                    Box::new(|rng: &mut R| property_insert_select(rng, env, &remaining_)),
+                    Box::new(|rng: &mut R| property_insert_values_select(rng, env, &remaining_)),
                 ),
                 (
                     remaining_.create / 2.0,
diff --git a/simulator/generation/query.rs b/simulator/generation/query.rs
index 1a0d167fc..3448bd434 100644
--- a/simulator/generation/query.rs
+++ b/simulator/generation/query.rs
@@ -1,13 +1,14 @@
 use crate::generation::table::{GTValue, LTValue};
 use crate::generation::{one_of, Arbitrary, ArbitraryFrom};
 
-use crate::model::query::{Create, Delete, Insert, Predicate, Query, Select};
+use crate::model::query::{Create, Delete, Distinctness, Insert, Predicate, Query, Select};
 use crate::model::table::{Table, Value};
+use crate::SimulatorEnv;
 use rand::seq::SliceRandom as _;
 use rand::Rng;
 
 use super::property::Remaining;
-use super::{frequency, pick};
+use super::{backtrack, frequency, pick};
 
 impl Arbitrary for Create {
     fn arbitrary<R: Rng>(rng: &mut R) -> Self {
@@ -17,81 +18,85 @@ impl Arbitrary for Create {
     }
 }
 
-impl ArbitraryFrom<&Vec<Table>> for Select {
-    fn arbitrary_from<R: Rng>(rng: &mut R, tables: &Vec<Table>) -> Self {
-        let table = pick(tables, rng);
+impl ArbitraryFrom<&SimulatorEnv> for Select {
+    fn arbitrary_from<R: Rng>(rng: &mut R, env: &SimulatorEnv) -> Self {
+        let table = pick(&env.tables, rng);
         Self {
             table: table.name.clone(),
             predicate: Predicate::arbitrary_from(rng, table),
             limit: Some(rng.gen_range(0..=1000)),
+            distinct: Distinctness::All,
         }
     }
 }
 
-impl ArbitraryFrom<&Vec<&Table>> for Select {
-    fn arbitrary_from<R: Rng>(rng: &mut R, tables: &Vec<&Table>) -> Self {
-        let table = pick(tables, rng);
-        Self {
-            table: table.name.clone(),
-            predicate: Predicate::arbitrary_from(rng, *table),
-            limit: Some(rng.gen_range(0..=1000)),
-        }
-    }
-}
-
-impl ArbitraryFrom<&Table> for Insert {
-    fn arbitrary_from<R: Rng>(rng: &mut R, table: &Table) -> Self {
-        let num_rows = rng.gen_range(1..10);
-        let values: Vec<Vec<Value>> = (0..num_rows)
-            .map(|_| {
-                table
-                    .columns
-                    .iter()
-                    .map(|c| Value::arbitrary_from(rng, &c.column_type))
-                    .collect()
+impl ArbitraryFrom<&SimulatorEnv> for Insert {
+    fn arbitrary_from<R: Rng>(rng: &mut R, env: &SimulatorEnv) -> Self {
+        let gen_values = |rng: &mut R| {
+            let table = pick(&env.tables, rng);
+            let num_rows = rng.gen_range(1..10);
+            let values: Vec<Vec<Value>> = (0..num_rows)
+                .map(|_| {
+                    table
+                        .columns
+                        .iter()
+                        .map(|c| Value::arbitrary_from(rng, &c.column_type))
+                        .collect()
+                })
+                .collect();
+            Some(Insert::Values {
+                table: table.name.clone(),
+                values,
             })
-            .collect();
-        Self {
-            table: table.name.clone(),
-            values,
-        }
-    }
-}
+        };
 
-impl ArbitraryFrom<&Table> for Delete {
-    fn arbitrary_from<R: Rng>(rng: &mut R, table: &Table) -> Self {
-        Self {
-            table: table.name.clone(),
-            predicate: Predicate::arbitrary_from(rng, table),
-        }
-    }
-}
+        let _gen_select = |rng: &mut R| {
+            // Find a non-empty table
+            let table = env.tables.iter().find(|t| !t.rows.is_empty());
+            if table.is_none() {
+                return None;
+            }
 
-impl ArbitraryFrom<&Table> for Query {
-    fn arbitrary_from<R: Rng>(rng: &mut R, table: &Table) -> Self {
-        frequency(
+            let select_table = table.unwrap();
+            let row = pick(&select_table.rows, rng);
+            let predicate = Predicate::arbitrary_from(rng, (select_table, row));
+            // Pick another table to insert into
+            let select = Select {
+                table: select_table.name.clone(),
+                predicate,
+                limit: None,
+                distinct: Distinctness::All,
+            };
+            let table = pick(&env.tables, rng);
+            Some(Insert::Select {
+                table: table.name.clone(),
+                select: Box::new(select),
+            })
+        };
+
+        backtrack(
             vec![
-                (1, Box::new(|rng| Self::Create(Create::arbitrary(rng)))),
-                (
-                    100,
-                    Box::new(|rng| Self::Select(Select::arbitrary_from(rng, &vec![table]))),
-                ),
-                (
-                    100,
-                    Box::new(|rng| Self::Insert(Insert::arbitrary_from(rng, table))),
-                ),
-                (
-                    0,
-                    Box::new(|rng| Self::Delete(Delete::arbitrary_from(rng, table))),
-                ),
+                (1, Box::new(|rng| gen_values(rng))),
+                // todo: test and enable this once `INSERT INTO <table> SELECT * FROM <table>` is supported
+                // (1, Box::new(|rng| gen_select(rng))),
             ],
             rng,
         )
     }
 }
 
-impl ArbitraryFrom<(&Table, &Remaining)> for Query {
-    fn arbitrary_from<R: Rng>(rng: &mut R, (table, remaining): (&Table, &Remaining)) -> Self {
+impl ArbitraryFrom<&SimulatorEnv> for Delete {
+    fn arbitrary_from<R: Rng>(rng: &mut R, env: &SimulatorEnv) -> Self {
+        let table = pick(&env.tables, rng);
+        Self {
+            table: table.name.clone(),
+            predicate: Predicate::arbitrary_from(rng, table),
+        }
+    }
+}
+
+impl ArbitraryFrom<(&SimulatorEnv, &Remaining)> for Query {
+    fn arbitrary_from<R: Rng>(rng: &mut R, (env, remaining): (&SimulatorEnv, &Remaining)) -> Self {
         frequency(
             vec![
                 (
@@ -100,15 +105,15 @@ impl ArbitraryFrom<(&Table, &Remaining)> for Query {
                 ),
                 (
                     remaining.read,
-                    Box::new(|rng| Self::Select(Select::arbitrary_from(rng, &vec![table]))),
+                    Box::new(|rng| Self::Select(Select::arbitrary_from(rng, env))),
                 ),
                 (
                     remaining.write,
-                    Box::new(|rng| Self::Insert(Insert::arbitrary_from(rng, table))),
+                    Box::new(|rng| Self::Insert(Insert::arbitrary_from(rng, env))),
                 ),
                 (
                     0.0,
-                    Box::new(|rng| Self::Delete(Delete::arbitrary_from(rng, table))),
+                    Box::new(|rng| Self::Delete(Delete::arbitrary_from(rng, env))),
                 ),
             ],
             rng,
diff --git a/simulator/model/query.rs b/simulator/model/query.rs
index 8c861edc4..bc97be0b2 100644
--- a/simulator/model/query.rs
+++ b/simulator/model/query.rs
@@ -101,7 +101,8 @@ impl Query {
         match self {
             Query::Create(_) => vec![],
             Query::Select(Select { table, .. })
-            | Query::Insert(Insert { table, .. })
+            | Query::Insert(Insert::Select { table, .. })
+            | Query::Insert(Insert::Values { table, .. })
             | Query::Delete(Delete { table, .. }) => vec![table.clone()],
         }
     }
@@ -109,12 +110,13 @@ impl Query {
         match self {
             Query::Create(Create { table }) => vec![table.name.clone()],
             Query::Select(Select { table, .. })
-            | Query::Insert(Insert { table, .. })
+            | Query::Insert(Insert::Select { table, .. })
+            | Query::Insert(Insert::Values { table, .. })
             | Query::Delete(Delete { table, .. }) => vec![table.clone()],
         }
     }
 
-    pub(crate) fn shadow(&self, env: &mut SimulatorEnv) {
+    pub(crate) fn shadow(&self, env: &mut SimulatorEnv) -> Vec<Vec<Value>> {
         match self {
             Query::Create(create) => create.shadow(env),
             Query::Insert(insert) => insert.shadow(env),
@@ -129,34 +131,110 @@ pub(crate) struct Create {
 }
 
 impl Create {
-    pub(crate) fn shadow(&self, env: &mut SimulatorEnv) {
+    pub(crate) fn shadow(&self, env: &mut SimulatorEnv) -> Vec<Vec<Value>> {
         if !env.tables.iter().any(|t| t.name == self.table.name) {
             env.tables.push(self.table.clone());
         }
+
+        vec![]
     }
 }
 
+impl Display for Create {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        write!(f, "CREATE TABLE {} (", self.table.name)?;
+
+        for (i, column) in self.table.columns.iter().enumerate() {
+            if i != 0 {
+                write!(f, ",")?;
+            }
+            write!(f, "{} {}", column.name, column.column_type)?;
+        }
+
+        write!(f, ")")
+    }
+}
+
+/// `SELECT` distinctness
+#[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
+pub enum Distinctness {
+    /// `DISTINCT`
+    Distinct,
+    /// `ALL`
+    All,
+}
+
 #[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
 pub(crate) struct Select {
     pub(crate) table: String,
     pub(crate) predicate: Predicate,
+    pub(crate) distinct: Distinctness,
     pub(crate) limit: Option<usize>,
 }
 
 impl Select {
-    pub(crate) fn shadow(&self, _env: &mut SimulatorEnv) {}
+    pub(crate) fn shadow(&self, env: &mut SimulatorEnv) -> Vec<Vec<Value>> {
+        let table = env.tables.iter().find(|t| t.name == self.table.as_str());
+        if let Some(table) = table {
+            table
+                .rows
+                .iter()
+                .filter(|row| self.predicate.test(row, table))
+                .cloned()
+                .collect()
+        } else {
+            vec![]
+        }
+    }
+}
+
+impl Display for Select {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        write!(
+            f,
+            "SELECT * FROM {} WHERE {}{}",
+            self.table,
+            self.predicate,
+            self.limit
+                .map_or("".to_string(), |l| format!(" LIMIT {}", l))
+        )
+    }
 }
 
 #[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
-pub(crate) struct Insert {
-    pub(crate) table: String,
-    pub(crate) values: Vec<Vec<Value>>,
+pub(crate) enum Insert {
+    Values {
+        table: String,
+        values: Vec<Vec<Value>>,
+    },
+    Select {
+        table: String,
+        select: Box<Select>,
+    },
 }
 
 impl Insert {
-    pub(crate) fn shadow(&self, env: &mut SimulatorEnv) {
-        if let Some(t) = env.tables.iter_mut().find(|t| t.name == self.table) {
-            t.rows.extend(self.values.clone());
+    pub(crate) fn shadow(&self, env: &mut SimulatorEnv) -> Vec<Vec<Value>> {
+        match self {
+            Insert::Values { table, values } => {
+                if let Some(t) = env.tables.iter_mut().find(|t| &t.name == table) {
+                    t.rows.extend(values.clone());
+                }
+            }
+            Insert::Select { table, select } => {
+                let rows = select.shadow(env);
+                if let Some(t) = env.tables.iter_mut().find(|t| &t.name == table) {
+                    t.rows.extend(rows);
+                }
+            }
+        }
+
+        vec![]
+    }
+
+    pub(crate) fn table(&self) -> &str {
+        match self {
+            Insert::Values { table, .. } | Insert::Select { table, .. } => table,
         }
     }
 }
@@ -168,7 +246,7 @@ pub(crate) struct Delete {
 }
 
 impl Delete {
-    pub(crate) fn shadow(&self, _env: &mut SimulatorEnv) {
+    pub(crate) fn shadow(&self, _env: &mut SimulatorEnv) -> Vec<Vec<Value>> {
         todo!()
     }
 }
@@ -176,30 +254,9 @@ impl Delete {
 impl Display for Query {
     fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
         match self {
-            Self::Create(Create { table }) => {
-                write!(f, "CREATE TABLE {} (", table.name)?;
-
-                for (i, column) in table.columns.iter().enumerate() {
-                    if i != 0 {
-                        write!(f, ",")?;
-                    }
-                    write!(f, "{} {}", column.name, column.column_type)?;
-                }
-
-                write!(f, ")")
-            }
-            Self::Select(Select {
-                table,
-                predicate: guard,
-                limit,
-            }) => write!(
-                f,
-                "SELECT * FROM {} WHERE {}{}",
-                table,
-                guard,
-                limit.map_or("".to_string(), |l| format!(" LIMIT {}", l))
-            ),
-            Self::Insert(Insert { table, values }) => {
+            Self::Create(create) => write!(f, "{}", create),
+            Self::Select(select) => write!(f, "{}", select),
+            Self::Insert(Insert::Values { table, values }) => {
                 write!(f, "INSERT INTO {} VALUES ", table)?;
                 for (i, row) in values.iter().enumerate() {
                     if i != 0 {
@@ -216,6 +273,10 @@ impl Display for Query {
                 }
                 Ok(())
             }
+            Self::Insert(Insert::Select { table, select }) => {
+                write!(f, "INSERT INTO {} ", table)?;
+                write!(f, "{}", select)
+            }
             Self::Delete(Delete {
                 table,
                 predicate: guard,
diff --git a/simulator/shrink/plan.rs b/simulator/shrink/plan.rs
index bc5eaa4d0..eed4fbd0f 100644
--- a/simulator/shrink/plan.rs
+++ b/simulator/shrink/plan.rs
@@ -30,7 +30,7 @@ impl InteractionPlan {
         for interaction in plan.plan.iter_mut() {
             if let Interactions::Property(p) = interaction {
                 match p {
-                    Property::InsertSelect { queries, .. }
+                    Property::InsertValuesSelect { queries, .. }
                     | Property::DoubleCreateFailure { queries, .. } => {
                         queries.clear();
                     }

From 849018464473309bf901440d89e30df82bfac56d Mon Sep 17 00:00:00 2001
From: alpaylan <alpkeles99@gmail.com>
Date: Sat, 1 Feb 2025 12:34:31 -0500
Subject: [PATCH 3/5] add LIKE generation for supporting  clauses

---
 Cargo.lock                    |  2 ++
 simulator/Cargo.toml          |  2 ++
 simulator/generation/mod.rs   | 12 +++++---
 simulator/generation/query.rs | 57 +++++++++++++++++++++++++----------
 simulator/generation/table.rs | 33 ++++++++++++++++++++
 simulator/model/query.rs      | 53 ++++++++++++++++++++++++++++----
 6 files changed, 133 insertions(+), 26 deletions(-)

diff --git a/Cargo.lock b/Cargo.lock
index 112c3793c..1fda860f2 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -1415,6 +1415,8 @@ dependencies = [
  "notify",
  "rand",
  "rand_chacha",
+ "regex",
+ "regex-syntax",
  "serde",
  "serde_json",
  "tempfile",
diff --git a/simulator/Cargo.toml b/simulator/Cargo.toml
index 43956e5e2..0c2a355a1 100644
--- a/simulator/Cargo.toml
+++ b/simulator/Cargo.toml
@@ -21,6 +21,8 @@ rand_chacha = "0.3.1"
 log = "0.4.20"
 tempfile = "3.0.7"
 env_logger = "0.10.1"
+regex = "1.11.1"
+regex-syntax = { version = "0.8.5", default-features = false, features = ["unicode"] }
 anarchist-readable-name-generator-lib = "0.1.2"
 clap = { version = "4.5", features = ["derive"] }
 serde = { version = "1.0", features = ["derive"] }
diff --git a/simulator/generation/mod.rs b/simulator/generation/mod.rs
index 71b8488be..ac7defd54 100644
--- a/simulator/generation/mod.rs
+++ b/simulator/generation/mod.rs
@@ -1,7 +1,4 @@
-use std::{
-    iter::Sum,
-    ops::SubAssign,
-};
+use std::{iter::Sum, ops::SubAssign};
 
 use anarchist_readable_name_generator_lib::readable_name_custom;
 use rand::{distributions::uniform::SampleUniform, Rng};
@@ -28,6 +25,13 @@ pub trait ArbitraryFrom<T> {
     fn arbitrary_from<R: Rng>(rng: &mut R, t: T) -> Self;
 }
 
+/// ArbitraryFromMaybe trait for fallibally generating random values from a given value
+pub trait ArbitraryFromMaybe<T> {
+    fn arbitrary_from_maybe<R: Rng>(rng: &mut R, t: T) -> Option<Self>
+    where
+        Self: Sized;
+}
+
 /// Frequency is a helper function for composing different generators with different frequency
 /// of occurences.
 /// The type signature for the `N` parameter is a bit complex, but it
diff --git a/simulator/generation/query.rs b/simulator/generation/query.rs
index 3448bd434..efafa6c05 100644
--- a/simulator/generation/query.rs
+++ b/simulator/generation/query.rs
@@ -8,7 +8,8 @@ use rand::seq::SliceRandom as _;
 use rand::Rng;
 
 use super::property::Remaining;
-use super::{backtrack, frequency, pick};
+use super::table::LikeValue;
+use super::{backtrack, frequency, pick, ArbitraryFromMaybe};
 
 impl Arbitrary for Create {
     fn arbitrary<R: Rng>(rng: &mut R) -> Self {
@@ -287,24 +288,48 @@ fn produce_true_predicate<R: Rng>(rng: &mut R, (t, row): (&Table, &Vec<Value>))
     let column_index = rng.gen_range(0..t.columns.len());
     let column = &t.columns[column_index];
     let value = &row[column_index];
-    one_of(
+    backtrack(
         vec![
-            Box::new(|_| Predicate::Eq(column.name.clone(), value.clone())),
-            Box::new(|rng| {
-                let v = loop {
+            (
+                1,
+                Box::new(|_| Some(Predicate::Eq(column.name.clone(), value.clone()))),
+            ),
+            (
+                1,
+                Box::new(|rng| {
                     let v = Value::arbitrary_from(rng, &column.column_type);
-                    if &v != value {
-                        break v;
+                    if &v == value {
+                        None
+                    } else {
+                        Some(Predicate::Neq(column.name.clone(), v))
                     }
-                };
-                Predicate::Neq(column.name.clone(), v)
-            }),
-            Box::new(|rng| {
-                Predicate::Gt(column.name.clone(), LTValue::arbitrary_from(rng, value).0)
-            }),
-            Box::new(|rng| {
-                Predicate::Lt(column.name.clone(), GTValue::arbitrary_from(rng, value).0)
-            }),
+                }),
+            ),
+            (
+                1,
+                Box::new(|rng| {
+                    Some(Predicate::Gt(
+                        column.name.clone(),
+                        LTValue::arbitrary_from(rng, value).0,
+                    ))
+                }),
+            ),
+            (
+                1,
+                Box::new(|rng| {
+                    Some(Predicate::Lt(
+                        column.name.clone(),
+                        GTValue::arbitrary_from(rng, value).0,
+                    ))
+                }),
+            ),
+            (
+                1,
+                Box::new(|rng| {
+                    LikeValue::arbitrary_from_maybe(rng, value)
+                        .map(|like| Predicate::Like(column.name.clone(), like.0))
+                }),
+            ),
         ],
         rng,
     )
diff --git a/simulator/generation/table.rs b/simulator/generation/table.rs
index 258cd1483..415b4b03a 100644
--- a/simulator/generation/table.rs
+++ b/simulator/generation/table.rs
@@ -3,6 +3,8 @@ use rand::Rng;
 use crate::generation::{gen_random_text, pick, readable_name_custom, Arbitrary, ArbitraryFrom};
 use crate::model::table::{Column, ColumnType, Name, Table, Value};
 
+use super::ArbitraryFromMaybe;
+
 impl Arbitrary for Name {
     fn arbitrary<R: Rng>(rng: &mut R) -> Self {
         let name = readable_name_custom("_", rng);
@@ -194,3 +196,34 @@ impl ArbitraryFrom<&Value> for GTValue {
         }
     }
 }
+
+pub(crate) struct LikeValue(pub(crate) String);
+
+impl ArbitraryFromMaybe<&Value> for LikeValue {
+    fn arbitrary_from_maybe<R: Rng>(rng: &mut R, value: &Value) -> Option<Self> {
+        match value {
+            Value::Text(t) => {
+                let mut t = t.chars().collect::<Vec<_>>();
+                // Remove a number of characters, either insert `_` for each character removed, or
+                // insert one `%` for the whole substring
+                let mut i = 0;
+                while i < t.len() {
+                    if rng.gen_bool(0.1) {
+                        t[i] = '_';
+                    } else if rng.gen_bool(0.05) {
+                        t[i] = '%';
+                        // skip a list of characters
+                        for _ in 0..rng.gen_range(0..=3.min(t.len() - i - 1)) {
+                            t.remove(i + 1);
+                        }
+                    }
+                    i += 1;
+                }
+                let index = rng.gen_range(0..t.len());
+                t.insert(index, '%');
+                Some(Self(t.into_iter().collect()))
+            }
+            _ => None,
+        }
+    }
+}
diff --git a/simulator/model/query.rs b/simulator/model/query.rs
index bc97be0b2..474a54f1e 100644
--- a/simulator/model/query.rs
+++ b/simulator/model/query.rs
@@ -1,5 +1,6 @@
 use std::fmt::Display;
 
+use regex::{Regex, RegexBuilder};
 use serde::{Deserialize, Serialize};
 
 use crate::{
@@ -9,12 +10,48 @@ use crate::{
 
 #[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
 pub(crate) enum Predicate {
-    And(Vec<Predicate>), // p1 AND p2 AND p3... AND pn
-    Or(Vec<Predicate>),  // p1 OR p2 OR p3... OR pn
-    Eq(String, Value),   // column = Value
-    Neq(String, Value),  // column != Value
-    Gt(String, Value),   // column > Value
-    Lt(String, Value),   // column < Value
+    And(Vec<Predicate>),  // p1 AND p2 AND p3... AND pn
+    Or(Vec<Predicate>),   // p1 OR p2 OR p3... OR pn
+    Eq(String, Value),    // column = Value
+    Neq(String, Value),   // column != Value
+    Gt(String, Value),    // column > Value
+    Lt(String, Value),    // column < Value
+    Like(String, String), // column LIKE Value
+}
+
+/// This function is a duplication of the exec_like function in core/vdbe/mod.rs at commit 9b9d5f9b4c9920e066ef1237c80878f4c3968524
+/// Any updates to the original function should be reflected here, otherwise the test will be incorrect.
+fn construct_like_regex(pattern: &str) -> Regex {
+    let mut regex_pattern = String::with_capacity(pattern.len() * 2);
+
+    regex_pattern.push('^');
+
+    for c in pattern.chars() {
+        match c {
+            '\\' => regex_pattern.push_str("\\\\"),
+            '%' => regex_pattern.push_str(".*"),
+            '_' => regex_pattern.push('.'),
+            ch => {
+                if regex_syntax::is_meta_character(c) {
+                    regex_pattern.push('\\');
+                }
+                regex_pattern.push(ch);
+            }
+        }
+    }
+
+    regex_pattern.push('$');
+
+    RegexBuilder::new(&regex_pattern)
+        .case_insensitive(true)
+        .dot_matches_new_line(true)
+        .build()
+        .unwrap()
+}
+
+fn exec_like(pattern: &str, text: &str) -> bool {
+    let re = construct_like_regex(pattern);
+    re.is_match(text)
 }
 
 impl Predicate {
@@ -43,6 +80,9 @@ impl Predicate {
             Predicate::Neq(column, value) => get_value(column) != Some(value),
             Predicate::Gt(column, value) => get_value(column).map(|v| v > value).unwrap_or(false),
             Predicate::Lt(column, value) => get_value(column).map(|v| v < value).unwrap_or(false),
+            Predicate::Like(column, value) => get_value(column)
+                .map(|v| exec_like(v.to_string().as_str(), value.as_str()))
+                .unwrap_or(false),
         }
     }
 }
@@ -83,6 +123,7 @@ impl Display for Predicate {
             Self::Neq(name, value) => write!(f, "{} != {}", name, value),
             Self::Gt(name, value) => write!(f, "{} > {}", name, value),
             Self::Lt(name, value) => write!(f, "{} < {}", name, value),
+            Self::Like(name, value) => write!(f, "{} LIKE '{}'", name, value),
         }
     }
 }

From 1b3d2004ccceb6402d398ad969b10ec5069f61f2 Mon Sep 17 00:00:00 2001
From: alpaylan <alpkeles99@gmail.com>
Date: Mon, 3 Feb 2025 09:57:04 -0500
Subject: [PATCH 4/5] change a potentially bug-hiding 'unwrap_or' call into
 'expect' for clear messaging

---
 simulator/generation/property.rs | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/simulator/generation/property.rs b/simulator/generation/property.rs
index 37c1fe744..860f11184 100644
--- a/simulator/generation/property.rs
+++ b/simulator/generation/property.rs
@@ -196,7 +196,7 @@ impl Property {
                     }),
                 });
 
-                let limit = select.limit.unwrap_or(0);
+                let limit = select.limit.expect("Property::SelectLimit without a LIMIT clause");
 
                 let assertion = Interaction::Assertion(Assertion {
                     message: "select query should respect the limit clause".to_string(),

From f8905d1bc0033221dbe5dadfa3806288c88473e4 Mon Sep 17 00:00:00 2001
From: alpaylan <alpkeles99@gmail.com>
Date: Wed, 5 Feb 2025 11:25:58 -0500
Subject: [PATCH 5/5] fix formatting

---
 simulator/generation/property.rs | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/simulator/generation/property.rs b/simulator/generation/property.rs
index 860f11184..3d4b9b4ce 100644
--- a/simulator/generation/property.rs
+++ b/simulator/generation/property.rs
@@ -196,7 +196,9 @@ impl Property {
                     }),
                 });
 
-                let limit = select.limit.expect("Property::SelectLimit without a LIMIT clause");
+                let limit = select
+                    .limit
+                    .expect("Property::SelectLimit without a LIMIT clause");
 
                 let assertion = Interaction::Assertion(Assertion {
                     message: "select query should respect the limit clause".to_string(),