metesynnada commented on PR #8020:
URL: 
https://github.com/apache/arrow-datafusion/pull/8020#issuecomment-1805943712

   Hi @korowa, as far as I understand, you reverted the build side indices per 
probe side index.
   
   ```
   Probe              Build                  Output
   ┌───────┬───────┐  ┌───────┬───────┐  ┌─────────┬─────────┬
   │  idx  │  key  │  │  idx  │  key  │  │probe_idx│build_idx│
   ├───────┼───────┤  ├───────┼───────┤  ├─────────┼─────────┼
   │   0   │   1   │  │   0   │   1   │  │    0    │    0    │
   │       │       │  │       │       │  │         │         │
   │   1   │   2   │  │   1   │   1   │  │    0    │    1    │
   │       │       │  │       │       │  │         │         │
   │   2   │   3   │  │   2   │   2   │  │    1    │    2    │
   │       │       │  │       │       │  │         │         │
   │   3   │   4   │  │   3   │   2   │  │    1    │    3    │
   │       │       │  │       │       │  │         │         │
   │   4   │   5   │  │   4   │   5   │  │    4    │    4    │
   └───────┴───────┘  └───────┴───────┘  └─────────┴─────────┴─
   ``` 
   Why do we construct the join result in a particular way? The order in which 
we construct the index is important because the build order is suffixed to the 
probe side order. For instance, if the build side order is a,b and the probe 
side order is x,y, then we expect the output order to be x,y,a,b.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to