[Improve][Connector-V2][HBase] Support configurable range scan boundary inclusion policies#10011
Conversation
…ry inclusion policies
…ry inclusion policies
…ry inclusion policies
| .withDescription( | ||
| "Whether to include the start row in the scan. Default is true (inclusive)."); | ||
|
|
||
| public static final Option<Boolean> END_ROW_INCLUSIVE = |
There was a problem hiding this comment.
In order to maintain consistency with the previous logic, i think the default value of end_row_inclusive should also be true.
There was a problem hiding this comment.
thanks @xiaochen-zhou suggest, to ensure that data can be read normally without duplication, we need to include only one side.
There was a problem hiding this comment.
hi,Will the data be duplicated when both END_ROW_INCLUSIVE and START_ROW_INCLUSIVE are true？
There was a problem hiding this comment.
I think the intended result of the configuration should be to control whether the boundaries are included. However, I believe that when both sides are set to false, it will result in shards like (start, 1), (1, 3), (3, end), which causes the loss of boundary data. Conversely, it will lead to duplicate boundary data. Instead of adjusting the boundaries of each split, we should only modify the inclusion settings of the start and end points
…`end_row_inclusive` parameters in HBase documentation.
|
@davidzollo @Carl-Zhou-CN I have added some doc, please take another look. |
Purpose of this pull request
Does this PR introduce any user-facing change?
How was this patch tested?
Check list
New License Guide