chore: add a setting to decide if table sample is deterministic (#16275)

Author: xudong963

src/query/expression/src/function.rs

@@ -117,6 +117,7 @@ pub struct FunctionContext {
     pub parse_datetime_ignore_remainder: bool,
     pub enable_dst_hour_fix: bool,
     pub enable_strict_datetime_parser: bool,
+    pub random_function_seed: bool,
 }
 
 impl Default for FunctionContext {
@@ -139,6 +140,7 @@ impl Default for FunctionContext {
             parse_datetime_ignore_remainder: false,
             enable_dst_hour_fix: false,
             enable_strict_datetime_parser: true,
+            random_function_seed: false,
         }
     }
 }

src/query/functions/src/scalars/other.rs

@@ -144,7 +144,11 @@ pub fn register(registry: &mut FunctionRegistry) {
             })
         },
         |ctx| {
-            let mut rng = rand::rngs::SmallRng::from_entropy();
+            let mut rng = if ctx.func_ctx.random_function_seed {
+                rand::rngs::SmallRng::seed_from_u64(1)
+            } else {
+                rand::rngs::SmallRng::from_entropy()
+            };
             let rand_nums = (0..ctx.num_rows)
                 .map(|_| rng.gen::<F64>())
                 .collect::<Vec<_>>();

src/query/service/src/sessions/query_ctx.rs

@@ -708,6 +708,7 @@ impl TableContext for QueryContext {
         let enable_dst_hour_fix = settings.get_enable_dst_hour_fix()?;
         let enable_strict_datetime_parser = settings.get_enable_strict_datetime_parser()?;
         let query_config = &GlobalConfig::instance().query;
+        let random_function_seed = settings.get_random_function_seed()?;
 
         Ok(FunctionContext {
             tz,
@@ -729,6 +730,7 @@ impl TableContext for QueryContext {
             parse_datetime_ignore_remainder,
             enable_dst_hour_fix,
             enable_strict_datetime_parser,
+            random_function_seed,
         })
     }
 

src/query/settings/src/settings_default.rs

@@ -824,6 +824,12 @@ impl DefaultSettings {
                     desc: "Format NULL as str in query api response",
                     mode: SettingMode::Both,
                     range: Some(SettingRange::Numeric(0..=1)),
+                }),
+                ("random_function_seed", DefaultSettingValue {
+                    value: UserSettingValue::UInt64(0),
+                    desc: "Seed for random function",
+                    mode: SettingMode::Both,
+                    range: Some(SettingRange::Numeric(0..=1)),
                 })
             ]);
 

src/query/settings/src/settings_getter_setter.rs

@@ -685,4 +685,8 @@ impl Settings {
     pub fn get_max_data_retention_period_in_days() -> u64 {
         DefaultSettings::data_retention_time_in_days_max()
     }
+
+    pub fn get_random_function_seed(&self) -> Result<bool> {
+        Ok(self.try_get_u64("random_function_seed")? == 1)
+    }
 }

tests/sqllogictests/suites/mode/standalone/explain/table_sample.test

@@ -1,42 +1,45 @@
 statement ok
 create or replace table t as select number as a from numbers(1000);
 
-## query T
-## explain select * from t sample row (10 rows);
-## ----
-## Filter
-## ├── output columns: [t.a (#0)]
-## ├── filters: [rand() <= 0.01]
-## ├── estimated rows: 200.00
-## └── TableScan
-##     ├── table: default.default.t
-##     ├── output columns: [a (#0)]
-##     ├── read rows: 1000
-##     ├── read size: 1.40 KiB
-##     ├── partitions total: 1
-##     ├── partitions scanned: 1
-##     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1>]
-##     ├── push downs: [filters: [rand() <= 0.01], limit: NONE]
-##     └── estimated rows: 1000.00
-## 
-## query T
-## explain select * from t sample row (99.1);
-## ----
-## Filter
-## ├── output columns: [t.a (#0)]
-## ├── filters: [rand() <= 0.991]
-## ├── estimated rows: 200.00
-## └── TableScan
-##     ├── table: default.default.t
-##     ├── output columns: [a (#0)]
-##     ├── read rows: 1000
-##     ├── read size: 1.40 KiB
-##     ├── partitions total: 1
-##     ├── partitions scanned: 1
-##     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1>]
-##     ├── push downs: [filters: [rand() <= 0.991], limit: NONE]
-##     └── estimated rows: 1000.00
-## 
+statement ok
+set random_function_seed = 1;
+
+query T
+explain select * from t sample row (10 rows);
+----
+Filter
+├── output columns: [t.a (#0)]
+├── filters: [rand() <= 0.01]
+├── estimated rows: 200.00
+└── TableScan
+    ├── table: default.default.t
+    ├── output columns: [a (#0)]
+    ├── read rows: 1000
+    ├── read size: 1.40 KiB
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1>]
+    ├── push downs: [filters: [rand() <= 0.01], limit: NONE]
+    └── estimated rows: 1000.00
+
+query T
+explain select * from t sample row (99.1);
+----
+Filter
+├── output columns: [t.a (#0)]
+├── filters: [rand() <= 0.991]
+├── estimated rows: 200.00
+└── TableScan
+    ├── table: default.default.t
+    ├── output columns: [a (#0)]
+    ├── read rows: 1000
+    ├── read size: 1.40 KiB
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1>]
+    ├── push downs: [filters: [rand() <= 0.991], limit: NONE]
+    └── estimated rows: 1000.00
+
 
 statement ok
 drop table t;
@@ -92,3 +95,6 @@ query
 select count(distinct number) < 10000000 from numbers(10000000) sample block(50);
 ----
 1
+
+statement ok
+set random_function_seed = 0;