`data-anonymization` `SelectFromDatabase` patch

README.md

Prevents SQL error caused by primary_key not being set

Previously the SelectFromDatabase strategy connects to the database, and extracts the values, when it's initalized. However, at this stage,the table isn't fully configured and so when it creates the source table, it doesn't set a primary key.

By changing it to lazily fetch, and cache, the values; we ensure that the source table will be correctly configured.

select_from_database.patch

diff --git a/lib/strategy/field/string/select_from_database.rb b/lib/strategy/field/string/select_from_database.rb
index 43bfc45..c7adef6 100644
--- a/lib/strategy/field/string/select_from_database.rb
+++ b/lib/strategy/field/string/select_from_database.rb
@@ -12,16 +12,23 @@ class SelectFromDatabase < SelectFromFile
         include Utils::Logging
 
         def initialize table_name, field_name, connection_spec
-          DataAnon::Utils::SourceDatabase.establish_connection connection_spec
-          source = Utils::SourceTable.create table_name, []
-          @values = source.select(field_name).distinct.collect { |record| record[field_name]}
-          logger.debug "For field strategy #{table_name}:#{field_name} using values #{@values} "
-
+          @table_name = table_name
+          @field_name = field_name
+          @connection_spec = connection_spec
         end
 
-      end
-
+        def anonymize field
+          @values ||= begin
+            DataAnon::Utils::SourceDatabase.establish_connection @connection_spec
+            source = Utils::SourceTable.create @table_name, []
+            values = source.select(@field_name).distinct.collect { |record| record[@field_name]}
+            logger.debug "For field strategy #{@table_name}:#{@field_name} using values #{values} "
+            values
+          end
 
+          super
+        end
+      end
     end
   end
 end
\ No newline at end of file