Github user zzeekk commented on the issue:
https://github.com/apache/spark/pull/7379
@IceMan81 Here is an abstract example of our workaround, building blocks as
additional equi-join conditions.
The task is to join Points (df1) on a line back to Segments of the same
line (df2): df1.id_line=df2.id_line and df1.position>=df2.position_from and
df1.position<df2.position_to
By building blocks for the position, we can add a more detailed equi-join
condition, but need to join twice and coalesce results to catch the case when
position_from is in a different block than position. Additional note: block
size (/10) must be adapted to expected data:
```
df1.as("df1")
.join( df2.as("df2a"), $"df1.id_line"===$"df2a.id_line" and
floor($"df1.position"/10)===floor($"df2a.position_from"/10)
and $"df1.position">=$"df2a.position_from" and
$"df1.position"<$"df2a.position_to", "left")
.join( df2.as("df2a"), $"df1.id_line"===$"df2a.id_line" and
floor($"df1.position"/10)===floor($"df2a.position_to"/10)
and $"df1.position">=$"df2a.position_from" and
$"df1.position"<$"df2a.position_to", "left")
```
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
