Use portable SQL

Author: georgestagg

src/naming.rs

@@ -88,6 +88,16 @@ pub const SOURCE_COLUMN: &str = concatcp!(GGSQL_PREFIX, "source", GGSQL_SUFFIX);
 /// Alias for schema extraction queries
 pub const SCHEMA_ALIAS: &str = concatcp!(GGSQL_SUFFIX, "schema", GGSQL_SUFFIX);
 
+// ============================================================================
+// Internal SQL Aliases
+// ============================================================================
+
+/// CTE name for recursive base counter in generate_series: `__ggsql_base__`
+pub const SERIES_BASE: &str = concatcp!(GGSQL_PREFIX, "base", GGSQL_SUFFIX);
+
+/// CTE name for sequence output from generate_series: `__ggsql_seq__`
+pub const SERIES_SEQ: &str = concatcp!(GGSQL_PREFIX, "seq", GGSQL_SUFFIX);
+
 // ============================================================================
 // Constructor Functions
 // ============================================================================

src/plot/layer/geom/boxplot.rs

@@ -9,7 +9,7 @@ use crate::{
         geom::types::get_column_name, DefaultAestheticValue, DefaultParam, DefaultParamValue,
         ParameterValue, StatResult,
     },
-    utils::{scalar_max, scalar_min},
+    utils::{sql_greatest, sql_least, sql_percentile},
     DataFrame, GgsqlError, Mappings, Result,
 };
 
@@ -164,8 +164,11 @@ fn stat_boxplot(
 
 fn boxplot_sql_compute_summary(from: &str, groups: &[String], value: &str, coef: &f64) -> String {
     let groups_str = groups.join(", ");
-    let lower_expr = scalar_max(&[&format!("q1 - {coef} * (q3 - q1)"), "min"]);
-    let upper_expr = scalar_min(&[&format!("q3 + {coef} * (q3 - q1)"), "max"]);
+    let lower_expr = sql_greatest(&[&format!("q1 - {coef} * (q3 - q1)"), "min"]);
+    let upper_expr = sql_least(&[&format!("q3 + {coef} * (q3 - q1)"), "max"]);
+    let q1 = sql_percentile(value, 0.25, from, groups);
+    let median = sql_percentile(value, 0.50, from, groups);
+    let q3 = sql_percentile(value, 0.75, from, groups);
     format!(
         "SELECT
           *,
@@ -176,18 +179,21 @@ fn boxplot_sql_compute_summary(from: &str, groups: &[String], value: &str, coef:
             {groups},
             MIN({value}) AS min,
             MAX({value}) AS max,
-            PERCENTILE_CONT(0.25) WITHIN GROUP (ORDER BY {value}) AS q1,
-            PERCENTILE_CONT(0.50) WITHIN GROUP (ORDER BY {value}) AS median,
-            PERCENTILE_CONT(0.75) WITHIN GROUP (ORDER BY {value}) AS q3
-          FROM ({from}) AS __ggsql_qt__
+            {q1} AS q1,
+            {median} AS median,
+            {q3} AS q3
+          FROM ({from}) AS _qt
           WHERE {value} IS NOT NULL
           GROUP BY {groups}
-        ) AS __ggsql_fn__",
+        ) AS _fn",
         lower_expr = lower_expr,
         upper_expr = upper_expr,
         groups = groups_str,
         value = value,
-        from = from
+        from = from,
+        q1 = q1,
+        median = median,
+        q3 = q3,
     )
 }
 
@@ -309,32 +315,37 @@ mod tests {
     fn test_sql_compute_summary_basic() {
         let groups = vec!["category".to_string()];
         let result = boxplot_sql_compute_summary("data", &groups, "value", &1.5);
-        assert!(result.contains("PERCENTILE_CONT(0.25) WITHIN GROUP (ORDER BY value)"));
-        assert!(result.contains("PERCENTILE_CONT(0.50) WITHIN GROUP (ORDER BY value)"));
-        assert!(result.contains("PERCENTILE_CONT(0.75) WITHIN GROUP (ORDER BY value)"));
+        assert!(result.contains("NTILE(4) OVER (ORDER BY value)"));
+        assert!(result.contains("AS q1"));
+        assert!(result.contains("AS median"));
+        assert!(result.contains("AS q3"));
         assert!(result.contains("MIN(value) AS min"));
         assert!(result.contains("MAX(value) AS max"));
         assert!(result.contains("WHERE value IS NOT NULL"));
         assert!(result.contains("GROUP BY category"));
-        assert!(result.contains("SELECT MAX(v) FROM (VALUES (q1 - 1.5"));
-        assert!(result.contains("SELECT MIN(v) FROM (VALUES (q3 + 1.5"));
+        assert!(result.contains("CASE WHEN (q1 - 1.5"));
+        assert!(result.contains("CASE WHEN (q3 + 1.5"));
     }
 
     #[test]
     fn test_sql_compute_summary_multiple_groups() {
         let groups = vec!["cat".to_string(), "region".to_string()];
         let result = boxplot_sql_compute_summary("tbl", &groups, "val", &1.5);
         assert!(result.contains("GROUP BY cat, region"));
-        assert!(result.contains("PERCENTILE_CONT(0.25) WITHIN GROUP (ORDER BY val)"));
+        assert!(result.contains("NTILE(4) OVER (ORDER BY val)"));
     }
 
     #[test]
     fn test_sql_compute_summary_custom_coef() {
         let groups = vec!["pos1".to_string()];
         let result = boxplot_sql_compute_summary("q", &groups, "pos2", &2.5);
         assert!(result.contains("2.5"));
-        assert!(result.contains("SELECT MAX(v) FROM (VALUES (q1 - 2.5 * (q3 - q1)), (min)) AS t(v)"));
-        assert!(result.contains("SELECT MIN(v) FROM (VALUES (q3 + 2.5 * (q3 - q1)), (max)) AS t(v)"));
+        assert!(
+            result.contains("(CASE WHEN (q1 - 2.5 * (q3 - q1)) >= (min) THEN (q1 - 2.5 * (q3 - q1)) ELSE (min) END)")
+        );
+        assert!(
+            result.contains("(CASE WHEN (q3 + 2.5 * (q3 - q1)) <= (max) THEN (q3 + 2.5 * (q3 - q1)) ELSE (max) END)")
+        );
     }
 
     #[test]
@@ -355,22 +366,27 @@ mod tests {
         let groups = vec!["category".to_string()];
         let result = boxplot_sql_compute_summary("SELECT * FROM sales", &groups, "price", &1.5);
 
-        let expected = r#"SELECT
+        let q1 = sql_percentile("price", 0.25, "SELECT * FROM sales", &groups);
+        let median = sql_percentile("price", 0.50, "SELECT * FROM sales", &groups);
+        let q3 = sql_percentile("price", 0.75, "SELECT * FROM sales", &groups);
+        let expected = format!(
+            r#"SELECT
           *,
-          (SELECT MAX(v) FROM (VALUES (q1 - 1.5 * (q3 - q1)), (min)) AS t(v)) AS lower,
-          (SELECT MIN(v) FROM (VALUES (q3 + 1.5 * (q3 - q1)), (max)) AS t(v)) AS upper
+          (CASE WHEN (q1 - 1.5 * (q3 - q1)) >= (min) THEN (q1 - 1.5 * (q3 - q1)) ELSE (min) END) AS lower,
+          (CASE WHEN (q3 + 1.5 * (q3 - q1)) <= (max) THEN (q3 + 1.5 * (q3 - q1)) ELSE (max) END) AS upper
         FROM (
           SELECT
             category,
             MIN(price) AS min,
             MAX(price) AS max,
-            PERCENTILE_CONT(0.25) WITHIN GROUP (ORDER BY price) AS q1,
-            PERCENTILE_CONT(0.50) WITHIN GROUP (ORDER BY price) AS median,
-            PERCENTILE_CONT(0.75) WITHIN GROUP (ORDER BY price) AS q3
-          FROM (SELECT * FROM sales) AS __ggsql_qt__
+            {q1} AS q1,
+            {median} AS median,
+            {q3} AS q3
+          FROM (SELECT * FROM sales) AS _qt
           WHERE price IS NOT NULL
           GROUP BY category
-        ) AS __ggsql_fn__"#;
+        ) AS _fn"#
+        );
 
         assert_eq!(result, expected);
     }
@@ -380,22 +396,27 @@ mod tests {
         let groups = vec!["region".to_string(), "product".to_string()];
         let result = boxplot_sql_compute_summary("SELECT * FROM data", &groups, "revenue", &1.5);
 
-        let expected = r#"SELECT
+        let q1 = sql_percentile("revenue", 0.25, "SELECT * FROM data", &groups);
+        let median = sql_percentile("revenue", 0.50, "SELECT * FROM data", &groups);
+        let q3 = sql_percentile("revenue", 0.75, "SELECT * FROM data", &groups);
+        let expected = format!(
+            r#"SELECT
           *,
-          (SELECT MAX(v) FROM (VALUES (q1 - 1.5 * (q3 - q1)), (min)) AS t(v)) AS lower,
-          (SELECT MIN(v) FROM (VALUES (q3 + 1.5 * (q3 - q1)), (max)) AS t(v)) AS upper
+          (CASE WHEN (q1 - 1.5 * (q3 - q1)) >= (min) THEN (q1 - 1.5 * (q3 - q1)) ELSE (min) END) AS lower,
+          (CASE WHEN (q3 + 1.5 * (q3 - q1)) <= (max) THEN (q3 + 1.5 * (q3 - q1)) ELSE (max) END) AS upper
         FROM (
           SELECT
             region, product,
             MIN(revenue) AS min,
             MAX(revenue) AS max,
-            PERCENTILE_CONT(0.25) WITHIN GROUP (ORDER BY revenue) AS q1,
-            PERCENTILE_CONT(0.50) WITHIN GROUP (ORDER BY revenue) AS median,
-            PERCENTILE_CONT(0.75) WITHIN GROUP (ORDER BY revenue) AS q3
-          FROM (SELECT * FROM data) AS __ggsql_qt__
+            {q1} AS q1,
+            {median} AS median,
+            {q3} AS q3
+          FROM (SELECT * FROM data) AS _qt
           WHERE revenue IS NOT NULL
           GROUP BY region, product
-        ) AS __ggsql_fn__"#;
+        ) AS _fn"#
+        );
 
         assert_eq!(result, expected);
     }