Storing a list of IDs as a VARCHAR/TEXT column can cause performance and data integrity problems. Querying against such a column would require using pattern-matching expressions. It is awkward and costly to join a comma-separated list to matching rows. This will make it harder to validate IDs. Think about what is the greatest number of entries this list must support? Instead of using a multi-valued attribute, consider storing it in a separate table, so that each individual value of that attribute occupies a separate row. Such an intersection table implements a many-to-many relationship between the two referenced tables. This will greatly simplify querying and validating the IDs.