Github user ijokarumawak commented on a diff in the pull request:
https://github.com/apache/nifi/pull/2518#discussion_r183994650
--- Diff:
nifi-nar-bundles/nifi-hbase-bundle/nifi-hbase-processors/src/main/java/org/apache/nifi/hbase/AbstractPutHBase.java
---
@@ -64,6 +65,18 @@
.addValidator(StandardValidators.NON_EMPTY_VALIDATOR)
.build();
+ protected static final PropertyDescriptor DEFAULT_VISIBILITY_STRING =
new PropertyDescriptor.Builder()
+ .name("hbase-default-vis-string")
+ .displayName("Default Visibility String")
+ .description("When using visibility labels, any value set in
this field will be applied to all cells that are written unless " +
+ "an attribute with the convention
\"visibility.COLUMN_FAMILY.COLUMN_QUALIFIER\" is present on the flowfile. If
this field " +
+ "is left blank, it will be assumed that no visibility
is to be set unless visibility-related attributes are set. NOTE: " +
+ "this configuration will have no effect on your data
if you have not enabled visibility labels in the HBase cluster.")
--- End diff --
If user would like to set visibility label per column family, how would
they do that? Default visibility string may be too broad but setting attributes
for each column qualifier can be error-prone as those configuration need to be
updated as new qualifier emerges. Isn't it beneficial to support column family
level default visibility? Probably by user defined dynamic property can be used.
---
