DMS/CDC of a single table

for Transformation rules where schema name is like 'BLAH' and Source table name is like '', rename is rule target schema? if so make sure Source table name is like '%'

separately tables in different tasks based on size or lob is a good practice.

but try to keep transactionally related tables in one task so that any referential integrity (foreign key, etc) is maintained on target.

It wont repopulate unless you

• restart the task thereby reloading the tables
• reload the table from table statistics (task has to be cdc only or full load and cdc)

For large tables, it is recommended to migrate them separately from other tables for better performance and control.

• Create separate DMS tasks - one task for the large tables only using selection rules to include just those tables, and another task for the remaining tables.
• Use table filtering - Include only the large tables in the existing task using selection rules. Then create a new task to migrate the remaining tables.
• Resume existing task for large tables - Pause the existing full migration task after it completes the smaller tables. Then modify the task to only include the large tables using selection rules, and resume it.
• The resume option allows reusing the same task without creating a new one. Make sure to monitor the tasks closely for any issues with the large tables. The AWS DMS documentation has more details on using selection rules, transformations, filtering tables, and resuming tasks.