Add support for distinct on for postgres #3472 #3513
Conversation
| @@ -574,6 +578,12 @@ assign(QueryCompiler.prototype, { | |||
| ); | |||
| }, | |||
|
|
|||
| distinctOn(value) { | |||
| throw new Error( | |||
| '.distinctOn() is currently only supported on PostgreSQL' | |||
There was a problem hiding this comment.
How difficult would it be to support equivalent MySQL syntax? https://stackoverflow.com/questions/17673457/converting-select-distinct-on-queries-from-postgresql-to-mysql
Also does that also work on redshift?
There was a problem hiding this comment.
@kibertoad I don't think it is a very good idea to emulate this feature in Knex because:
- It introduces more complexity (both in implementation, and affects runtime performance)
- I don't see a straightforward emulation that is generically applicable
For instance, consider the last example in that answer, if we add one more column (col5) to the order by then the solution will need more modification because we would need to find the minimum of col5 among all rows that have same value of col4 potentially introducing another level of nesting.
Redshift doesn't support distinct on either.
lorefnon
force-pushed
the
#3472
branch
2 times, most recently
from
6142b74 to
5a4a12c
Compare
lorefnon
changed the title
|
@lorefnon Can you also add documentation for it? |
Interface as provided by knex#3513
Interface as provided by knex#3513
Interface as provided by knex#3513
No description provided.