Shuffling data stored in columnar tables improves data storage security, particularly when used in conjunction with other security operations, such as tokenization and cryptography. A data table is accessed, and pointer values of at least one column of the accessed table are shuffled, generating a protected table. An index table mapping index values to the shuffled pointer values is generated, allowing a user with access to both the protected table and the index table to generate the original table. Without both tables, users are only able to see either the shuffled data or the index values. Example shuffling methods include, but are not limited to, random shuffling, grouped shuffling, sorting by column value, and sorting by index value.