Primitive keys: Support INT, BIGINT, DOUBLE and STRING in JOIN
#4094
Assignees
Milestone
Comments
big-andy-coates
added a commit
to big-andy-coates/ksql
that referenced
this issue
Dec 10, 2019
Fixes: confluentinc#4092 WIP: This commit gets `PARTITION BY` clauses working with primitive key types. However, it does disable a couple of join until confluentinc#4094 has been completed. BREAKING CHANGE: A `PARTITION BY` now changes the SQL type of `ROWKEY` in the output schema of a query. For example, consider: ```sql CREATE STREAM INPUT (ROWKEY STRING KEY, ID INT) WITH (...); CREATE STREAM OUTPUT AS SELECT ROWKEY AS NAME FROM INPUT PARTITION BY ID; ``` Previously, the above would have resulted in an output schema of `ROWKEY STRING KEY, NAME STRING`, where `ROWKEY` would have stored the string representation of the integer from the `ID` column. With this commit the output schema will be `ROWKEY INT KEY, NAME STRING`.
2 tasks
big-andy-coates
added a commit
that referenced
this issue
Dec 10, 2019
* chore: partition-by primitive key support Fixes: #4092 WIP: This commit gets `PARTITION BY` clauses working with primitive key types. However, it does disable a couple of join until #4094 has been completed. BREAKING CHANGE: A `PARTITION BY` now changes the SQL type of `ROWKEY` in the output schema of a query. For example, consider: ```sql CREATE STREAM INPUT (ROWKEY STRING KEY, ID INT) WITH (...); CREATE STREAM OUTPUT AS SELECT ROWKEY AS NAME FROM INPUT PARTITION BY ID; ``` Previously, the above would have resulted in an output schema of `ROWKEY STRING KEY, NAME STRING`, where `ROWKEY` would have stored the string representation of the integer from the `ID` column. With this commit the output schema will be `ROWKEY INT KEY, NAME STRING`.
2 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
No description provided.
The text was updated successfully, but these errors were encountered: