[ 
https://issues.apache.org/jira/browse/RANGER-4793?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17848009#comment-17848009
 ] 

Madhan Neethiraj commented on RANGER-4793:
------------------------------------------

[~zhongwei11] - wildcard is currently not enabled for masking and row-filter 
policies to reduce the chances of setting up incompatible policies. For example:
 - mask type HASH is only applicable for columns having string values. With use 
of wildcard, it is easily possible for this mask type to be setup for columns 
having other datatypes. Similarly for other mask types that are valid only for 
specific datatypes
 - row-filter expressions typically refer to columns in the table on which the 
filter is applied. With use of wildcard, it is easily possible to setup a 
filter on a table that doesn't include the referenced column

Over the years, there have been many asks to enable wildcards for masking and 
row-filter policies - to reduce number of policies hence making it easier to 
manage. This certainly requires the policy author to setup policies to avoid 
incompatible masking and filtering.

{colType\} macro introduced in RANGER-4650 
([https://reviews.apache.org/r/74834/]) can help setup datatype specific mask 
type.

Perhaps introducing a new policy condition like COLUMNS_EXIST('state') can help 
with row-filter as well. This should be explored further.

About the changes in this patch, in addition to supporting wildcards, it will 
help to enable multiple columns/tables to be listed in policies. I suggest to 
remove attributes other than "name" in #353 - #378 and #443 - #462 - as shown 
below, which would result in values defined earlier for these resources to be 
used in dataMaskDef and rowFilterDef:

{noformat}
"dataMaskDef": {
  ...
  "resources": [
    { "name": "database" },
    { "name": "table" },
    { "name": "column" }
  ],
},
"rowFilterDef": {
  ...
  "resources": [
    { "name": "database" },
    { "name": "table" }
  ]
}
{noformat}


> Make HADOOPSQL Masking and Row Level Filter policies support wildcards by 
> default
> ---------------------------------------------------------------------------------
>
>                 Key: RANGER-4793
>                 URL: https://issues.apache.org/jira/browse/RANGER-4793
>             Project: Ranger
>          Issue Type: Improvement
>          Components: plugins
>    Affects Versions: 2.2.0, 2.3.0, 2.4.0
>            Reporter: wangzhongwei
>            Assignee: wangzhongwei
>            Priority: Major
>         Attachments: RANGER-4793.patch
>
>
>  Now ,in Access Policies ,input of database,table,and column support 
> wildcards,while  Masking and Row Level Filter do not work when using 
> wildcards. 



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to