Hey Team I have a kedro pipeline which saves its last 5 node

Question

Hey Team I have a kedro pipeline which saves its last 5 nodes data directly to database using kedroSQLTableDataSet Is there a way to implement rollback of all data that was saved in case any one of the data fails during insertion appending to the database I saw that we can create a custom dataset to do a bulk insertion operation within a transaction Please let me know what would be the best approach to do this Thanks

Sneha Kumari · Answer

< Priyanka Patil> fyi

Nok Lam Chan · Answer

How does `kedroSQLTableDataSet` different from `SQLTableDataSet`

datajoely · Answer

if you are using the default pandas dataset available in Kedro remember the following `save args` are available to help manage this sort of things

datajoely · Answer

If you really need to create your own dataset you can copy our implementation and add this sort of pattern to the `save` method <https stackoverflow com a 63442600 2010808>