For every importer / record combo there should be exactly one entry (#…

…921)

* for every importer record combo there should be an entry. no longer reuse entries from previous importers

* entry identifier is no longer unique and needs a scoped index

app/parsers/bulkrax/application_parser.rb

@@ -305,11 +305,15 @@ def new_entry(entryclass, type)
     end
 
     def find_or_create_entry(entryclass, identifier, type, raw_metadata = nil)
-      entry = entryclass.where(
+      # limit entry search to just this importer or exporter. Don't go moving them
+      entry = importerexporter.entries.where(
+        identifier: identifier
+      ).first
+      entry ||= entryclass.new(
         importerexporter_id: importerexporter.id,
         importerexporter_type: type,
         identifier: identifier
-      ).first_or_create!
+      )
       entry.raw_metadata = raw_metadata
       entry.save!
       entry

db/migrate/20240209070952_update_identifier_index.rb

@@ -0,0 +1,6 @@
+class UpdateIdentifierIndex < ActiveRecord::Migration[5.2]
+  def change
+    remove_index :bulkrax_entries, :identifier if index_exists?(:bulkrax_entries, :identifier )
+    add_index :bulkrax_entries, [:identifier, :importerexporter_id, :importerexporter_type], name: 'bulkrax_identifier_idx' unless index_exists?(:bulkrax_entries, [:identifier, :importerexporter_id, :importerexporter_type], name: 'bulkrax_identifier_idx')
+  end
+end

spec/test_app/db/schema.rb

@@ -10,7 +10,7 @@
 #
 # It's strongly recommended that you check this file into your version control system.
 
-ActiveRecord::Schema.define(version: 2024_02_08_005801) do
+ActiveRecord::Schema.define(version: 2024_02_09_070952) do
 
   create_table "accounts", force: :cascade do |t|
     t.string "name"
@@ -42,7 +42,7 @@
     t.string "importerexporter_type", default: "Bulkrax::Importer"
     t.integer "import_attempts", default: 0
     t.string "status_message", default: "Pending"
-    t.index ["identifier"], name: "index_bulkrax_entries_on_identifier"
+    t.index ["identifier", "importerexporter_id", "importerexporter_type"], name: "bulkrax_identifier_idx"
     t.index ["importerexporter_id", "importerexporter_type"], name: "bulkrax_entries_importerexporter_idx"
     t.index ["type"], name: "index_bulkrax_entries_on_type"
   end