From 2eee29c0bf17e1184880dae5804cc37b1b4b5aa2 Mon Sep 17 00:00:00 2001
From: t8y2 <1156263951@qq.com>
Date: Tue, 18 Aug 2026 06:32:44 +0000
Subject: [PATCH] fix(grid): deduplicate copied SQL IN tuples

Closes #6542
---
 .../dbx-core/src/data_grid_extractors/sql.rs  | 26 ++++++----
 .../src/data_grid_extractors/tests.rs         | 49 +++++++++++++++++++
 2 files changed, 66 insertions(+), 9 deletions(-)

diff --git a/crates/dbx-core/src/data_grid_extractors/sql.rs b/crates/dbx-core/src/data_grid_extractors/sql.rs
index 6b9f994a26..f387c27a1b 100644
--- a/crates/dbx-core/src/data_grid_extractors/sql.rs
+++ b/crates/dbx-core/src/data_grid_extractors/sql.rs
@@ -10,7 +10,7 @@ use crate::data_grid_sql::{
     DataGridCopyUpdateStatementOptions, DataGridTableMeta,
 };
 use serde_json::Value;
-use std::collections::HashSet;
+use std::collections::{hash_map::Entry, HashMap, HashSet};
 use std::io::Write;
 
 pub(super) fn write_sql_in_list(
@@ -18,23 +18,31 @@ pub(super) fn write_sql_in_list(
     output: &mut dyn Write,
 ) -> Result<(), DataGridExtractError> {
     write_bytes(output, b"(")?;
-    for (row_index, row) in context.request.rows.iter().enumerate() {
-        if row_index > 0 {
-            write_bytes(output, b", ")?;
-        }
+    let mut seen_rows = HashMap::<Vec<u8>, ()>::new();
+    let mut has_output = false;
+    for row in &context.request.rows {
+        let mut tuple = Vec::new();
         if context.selected_columns.len() > 1 {
-            write_bytes(output, b"(")?;
+            tuple.push(b'(');
         }
         for (column_index, source_index) in context.selected_source_indexes.iter().enumerate() {
             if column_index > 0 {
-                write_bytes(output, b", ")?;
+                tuple.extend_from_slice(b", ");
             }
             let value = &row[*source_index];
             let info = context.selected_column_info[column_index];
-            write_bytes(output, format_grid_sql_literal(value, context.request.database_type, info).as_bytes())?;
+            tuple.extend_from_slice(format_grid_sql_literal(value, context.request.database_type, info).as_bytes());
         }
         if context.selected_columns.len() > 1 {
-            write_bytes(output, b")")?;
+            tuple.push(b')');
+        }
+        if let Entry::Vacant(entry) = seen_rows.entry(tuple) {
+            if has_output {
+                write_bytes(output, b", ")?;
+            }
+            write_bytes(output, entry.key())?;
+            entry.insert(());
+            has_output = true;
         }
     }
     write_bytes(output, b")")
diff --git a/crates/dbx-core/src/data_grid_extractors/tests.rs b/crates/dbx-core/src/data_grid_extractors/tests.rs
index 5ddbd849a1..569a0fdc71 100644
--- a/crates/dbx-core/src/data_grid_extractors/tests.rs
+++ b/crates/dbx-core/src/data_grid_extractors/tests.rs
@@ -187,6 +187,55 @@ fn extracts_multi_column_sql_in_as_row_value_tuples() {
     assert_eq!(result.text, "((1, 'Ada'), (2, 'Grace, Hopper'))");
 }
 
+#[test]
+fn sql_in_list_deduplicates_single_selected_values() {
+    let mut request = request(DataGridExtractorId::SqlInList);
+    request.selected_column_indexes = vec![0];
+    request.rows = vec![
+        vec![json!(1), json!("first")],
+        vec![json!(1), json!("unselected value differs")],
+        vec![json!(2), json!("last")],
+    ];
+
+    let result = extract_data_grid_selection(request).expect("SQL IN extraction");
+
+    assert_eq!(result.text, "(1, 2)");
+}
+
+#[test]
+fn sql_in_list_deduplicates_multi_column_tuples_in_first_seen_order() {
+    let mut request = request(DataGridExtractorId::SqlInList);
+    request.rows = vec![
+        vec![json!(2), json!("Grace")],
+        vec![json!(1), json!("Ada")],
+        vec![json!(2), json!("Grace")],
+        vec![json!(1), json!("Lovelace")],
+        vec![json!(1), json!("Ada")],
+    ];
+
+    let result = extract_data_grid_selection(request).expect("SQL IN extraction");
+
+    assert_eq!(result.text, "((2, 'Grace'), (1, 'Ada'), (1, 'Lovelace'))");
+}
+
+#[test]
+fn sql_in_list_deduplicates_only_identical_rendered_literals() {
+    let mut request = request(DataGridExtractorId::SqlInList);
+    request.selected_column_indexes = vec![0];
+    request.rows = vec![
+        vec![json!(1)],
+        vec![json!("1")],
+        vec![Value::Null],
+        vec![Value::Null],
+        vec![json!("O'Reilly")],
+        vec![json!("O'Reilly")],
+    ];
+
+    let result = extract_data_grid_selection(request).expect("SQL IN extraction");
+
+    assert_eq!(result.text, "(1, '1', NULL, 'O''Reilly')");
+}
+
 #[test]
 fn preserves_duplicate_json_columns_without_overwriting_values() {
     let mut request = request(DataGridExtractorId::Json);