From f0b5d57998539611bf4e2ef88832e2ff6899db70 Mon Sep 17 00:00:00 2001
From: dal <dallin@buster.so>
Date: Wed, 23 Apr 2025 11:55:02 -0600
Subject: [PATCH 2/2] hotfix - deploy replacing dataset id

---
 .../rest/routes/datasets/deploy_datasets.rs   | 46 +++++++++++++++++--
 1 file changed, 42 insertions(+), 4 deletions(-)

diff --git a/api/server/src/routes/rest/routes/datasets/deploy_datasets.rs b/api/server/src/routes/rest/routes/datasets/deploy_datasets.rs
index ee6c07f3d..7ce2a437a 100644
--- a/api/server/src/routes/rest/routes/datasets/deploy_datasets.rs
+++ b/api/server/src/routes/rest/routes/datasets/deploy_datasets.rs
@@ -328,6 +328,32 @@ async fn deploy_datasets_handler(
             }
         };
 
+        // ---- Fetch Existing Dataset IDs for this data source ----
+        let request_db_names: Vec<String> = group.iter().map(|req| req.name.clone()).collect();
+        let existing_dataset_ids: HashMap<String, Uuid> = match datasets::table
+            .filter(datasets::data_source_id.eq(data_source.id))
+            .filter(datasets::database_name.eq_any(&request_db_names))
+            // .filter(datasets::deleted_at.is_null()) // Keep this commented if we want to reuse IDs of soft-deleted datasets
+            .select((datasets::database_name, datasets::id))
+            .load::<(String, Uuid)>(&mut conn)
+            .await {
+                Ok(ids) => ids.into_iter().collect(),
+                Err(e) => {
+                    tracing::error!(
+                        "Failed to retrieve existing dataset IDs for data source '{}': {}",
+                        data_source_name,
+                        e
+                    );
+                    // If we can't fetch existing IDs, we might proceed with new IDs,
+                    // or fail the whole batch for this data source. Failing is safer.
+                    return Err(anyhow::anyhow!(
+                        "Failed to retrieve existing dataset IDs: {}",
+                        e
+                    ));
+                }
+            };
+        // ---- End Fetch Existing IDs ----
+
         // Process all requests in the group directly
         let mut datasets_to_upsert_map: HashMap<(String, Uuid), Dataset> = HashMap::new();
         let mut columns_to_upsert_map: HashMap<String, Vec<DatasetColumn>> = HashMap::new();
@@ -344,14 +370,27 @@ async fn deploy_datasets_handler(
             results.push(validation); // Add to results now
 
             let now = Utc::now();
-            let dataset_id = req.id.unwrap_or_else(Uuid::new_v4);
+
+            // ---- Determine Dataset ID: Prioritize existing ID ----
+            let dataset_id = match existing_dataset_ids.get(&req.name) {
+                Some(existing_id) => {
+                    tracing::debug!("Using existing dataset ID {} for '{}'", existing_id, req.name);
+                    *existing_id
+                }
+                None => {
+                    let new_id = req.id.unwrap_or_else(Uuid::new_v4);
+                    tracing::debug!("Using new/request dataset ID {} for '{}'", new_id, req.name);
+                    new_id
+                }
+            };
+            // ---- End Determine Dataset ID ----
 
             // Prepare Dataset for upsert
             let dataset = Dataset {
-                id: dataset_id,
+                id: dataset_id, // Use the determined ID (existing or new/request)
                 name: req.name.clone(),
                 data_source_id: data_source.id, // Use fetched data source ID
-                created_at: now,
+                created_at: now, // Will be ignored by upsert if row exists
                 updated_at: now,
                 database_name: req.name.clone(), // Use model name as database_name
                 when_to_use: Some(req.description.clone()),
@@ -434,7 +473,6 @@ async fn deploy_datasets_handler(
                 .on_conflict((datasets::database_name, datasets::data_source_id))
                 .do_update()
                 .set((
-                    datasets::id.eq(excluded(datasets::id)), // Ensure ID is updated on conflict if new one generated
                     datasets::name.eq(excluded(datasets::name)),
                     datasets::updated_at.eq(now), // Use current time for update
                     datasets::updated_by.eq(excluded(datasets::updated_by)),