Using Row-Level Security (RLS) to Restrict Access to a Dataset - Amazon QuickSight

Using Row-Level Security (RLS) to Restrict Access to a Dataset

In the Enterprise edition of Amazon QuickSight, you can restrict access to a dataset by configuring row-level security (RLS) on it. You can do this before or after you have shared the dataset. Only the people you shared with can see any of the data. By adding row-level security, you can further control their access.

To do this, you create a query or file that has one column named UserName, GroupName, or both. Alternatively, you can create a query or file that has one column named UserARN, GroupARN, or both. You can think of this as adding a rule for that user or group. Then you can add one column to the query or file for each field that you want to grant or restrict access to. For each user or group name that you add, you add the values for each field. You can use NULL (no value) to mean all values. To see examples of dataset rules, see Creating Dataset Rules for Row-Level Security.

To apply the dataset rules, you add the rules as a permissions dataset to your dataset. Keep in mind the following points:

  • The permissions dataset can't contain duplicate values. Duplicates are ignored when evaluating how to apply the rules.

  • Each user or group specified can see only the rows that match the field values in the dataset rules.

  • If you add a rule for a user or group and leave all other columns with no value (NULL), you grant them access to all the data.

  • If you don't add a rule for a user or group, that user or group can't see any of the data.

  • The full set of rule records that are applied per user must not exceed 999. This applies to the total number of rules that are directly assigned to a user name plus any rules that are assigned to the user through group names.

  • If a field includes a comma (,) Amazon QuickSight treats each word separated from another by a comma as an individual value in the filter. For example, in ('AWS', 'INC'), AWS,INC is considered as two strings: AWS and INC. To filter with AWS,INC, wrap the string with double quotation marks in the permissions dataset.

Amazon QuickSight treats spaces as literal values. If you have a space in a field that you are restricting, the dataset rule applies to those rows. Amazon QuickSight treats both NULLs and blanks (empty strings “”) as "no value". A NULL is an empty field value.

Depending on what data source your dataset is coming from, you can configure a direct query to access a table of permissions. Terms with spaces inside them don't need to be delimited with quotes. If you use a direct query, you can easily change the query in the original data source.

Or you can upload dataset rules from a text file or spreadsheet. If you are using a comma-separated value (CSV) file, don't include any spaces on the given line. Terms with spaces inside them need to be delimited with quotes. If you use dataset rules that are file-based, apply any changes by overwriting the existing rules in the dataset's permissions settings.

Data sets that are restricted are marked with the word RESTRICTED in the Datasets screen.

Row-level security only works for fields containing textual data (string, char, varchar, and so on). It doesn't currently work for dates or numeric fields. Anomaly detection is not supported for datasets that use row-level security (RLS).

Creating Dataset Rules for Row-Level Security

Use the following procedure to create a permissions file or query to use as dataset rules.

To create a permissions files or query to use as dataset rules

  1. Create a file or a query that contains the dataset rules (permissions) for row-level security.

    It doesn't matter what order the fields are in. However, all the fields are case-sensitive. They must exactly match the field names and values.

    The structure should look similar to one of the following. You must have at least one field that identifies either users or groups. You can include both, but only one is required, and only one is used at a time. The field you use for users or groups can have any name you choose.


    If you are specifying groups, use only Amazon QuickSight groups or Microsoft AD groups.

    The following example shows a table with groups.

    GroupName Region Segment
    EMEA-Sales EMEA Enterprise,SMB,Startup
    US-Sales US Enterprise
    US-Sales US SMB, Startup
    US-Sales US Startup
    APAC-Sales APAC Enterprise,SMB
    APAC-Sales APAC Enterprise,Startup

    The following example shows a table with user names.

    UserName Region Segment
    AlejandroRosalez EMEA Enterprise,SMB,Startup
    MarthaRivera US Enterprise
    NikhilJayashankar US SMB,Startup
    PauloSantos US Startup
    SaanviSarkar APAC Enterprise,SMB
    ZhangWei APAC Enterprise,Startup

    The following example shows a table with user and group ARNs.

    UserARN GroupARN Region
    arn:aws:quicksight:us-east-1:123456789012:user/Bob arn:aws:quicksight:us-east-1:123456789012:group/group-1 APAC
    arn:aws:quicksight:us-east-1:123456789012:user/Sam arn:aws:quicksight:us-east-1:123456789012:group/group-2 US

    Or if you prefer to use a .csv file, the structure should look similar to one of the following.

    UserName,Region,Segment AlejandroRosalez,EMEA,"Enterprise,SMB,Startup" MarthaRivera,US,Enterprise NikhilJayashankars,US,SMB PauloSantos,US,Startup SaanviSarkar,APAC,"SMB,Startup","","" ZhangWei,APAC-Sales,"Enterprise,Startup"
    GroupName,Region,Segment EMEA-Sales,EMEA,"Enterprise,SMB,Startup" US-Sales,US,Enterprise US-Sales,US,SMB US-Sales,US,Startup APAC-Sales,APAC,"SMB,Startup" Corporate-Reporting,"","" APAC-Sales,APAC,"Enterprise,Startup"
    UserARN,GroupARN,Region arn:aws:quicksight:us-east-1:123456789012:user/Bob,arn:aws:quicksight:us-east-1:123456789012:group/group-1,APAC arn:aws:quicksight:us-east-1:123456789012:user/Sam,arn:aws:quicksight:us-east-1:123456789012:group/group-2,US

    Following is a SQL example.

    /* for users*/ select User as UserName, Region, Segment from tps-permissions; /* for groups*/ select Group as GroupName, Region, Segment from tps-permissions;
  2. Create a dataset for the dataset rules. To make sure that you can easily find it, give it a meaningful name, for example Permissions-Sales-Pipeline.

Creating Row-Level Security

Use the following procedure to apply row-level security (RLS) by using a file or query as a dataset that contains the rules for permissions.

To apply row-level security by using a file or query

  1. Confirm that you have added your rules as a new dataset. If you added them, but don't see them under the list of datasets, refresh the screen.

  2. On the Datasets page, choose the dataset, and then choose Permissions.

  3. From the list of datasets, choose your permissions dataset.

    If your permissions dataset doesn't appear on this screen, return to your datasets, and refresh the page.

  4. Choose the permissions policy. Each dataset has only one active permissions dataset. If you try to add a second permissions dataset, it overwrites the existing one.


    Some restrictions apply to NULL and empty string values when working with row-level security.

    • If your dataset has NULL values or empty strings (“”) in the restricted fields, these rows are ignored when the restrictions are applied.

    • Inside the permissions dataset, NULL values and empty strings are treated the same. For more information, see the following table.

    • To prevent accidentally exposing sensitive information, QuickSight skips empty RLS rules that grant access to everyone. An empty RLS rule occurs when all columns of a row have no value. QuickSight RLS treats NULL, empty strings (""), or empty comma separated strings (e.g. ",,,") as no value.

      • After skipping empty rules, other non-empty RLS rules still apply.

      • If a permission dataset has only empty rules and all of them were skipped, no one has access to any data restricted by this permission dataset.

    Rules for UserName, GroupName, Region, Segment Granted Access
    AlejandroRosalez,EMEA-Sales,EMEA,"Enterprise,SMB,Startup" Sees all EMEA Enterprise, SMB, and Startup,Corporate-Reporting,"","" Sees all rows
    User or group has no entry Sees no rows
    “”,“”,“”,“” Skipped; sees no rows if all other rules are empty.
    NULL,“”,“”,NULL Skipped; sees no rows if all other rules are empty.

    Anyone you shared your dashboard with can see all the data in it, unless the dataset is restricted by dataset rules.

  5. To save your changes, choose Apply data set. Then, on the Confirm: saving data set rules screen, choose Apply data set. Changes in permissions apply immediately to existing users.

  6. (Optional) To remove permissions, first remove the dataset rules from the dataset.

    Make certain the dataset rules are removed. Then, choose the permissions dataset and choose Remove data set.

    To overwrite permissions, choose a new permissions dataset and apply it. You can reuse the same dataset name, but you need to apply the new permissions in the Permissions screen to make these permissions active. SQL queries dynamically update, so these can be managed outside of Amazon QuickSight. For queries, the permissions are updated when the direct query cache is automatically refreshed.

If you delete a file-based permissions dataset before you remove it from the target dataset, restricted users can't access the dataset. While the dataset is in this state, it remains marked as RESTRICTED. However, when you view Permissions for that dataset, you can see that it has no selected dataset rules. To fix this, you can specify new dataset rules. Creating a dataset with the same name is not enough to fix this. You must choose the new permissions dataset in the Permissions screen. This restriction doesn't apply to direct SQL queries.