Sign up to take part
Registered users can ask their own questions, contribute to discussions, and be part of the Community!
Registered users can ask their own questions, contribute to discussions, and be part of the Community!
New to dataiku & running into trouble trying to find a way to get the unjoined records.
Something similar to L&R anchor in an Alteryx join tool (https://help.alteryx.com/20223/designer/join-tool)
Hi @XiShi ,
Great question! You can get unjoined rows (also called an anti-join) using the Join recipe. Here's how:
Similarly, you could also opt for a full outer join like @tgb417 suggested and follow it with a split recipe to create three datasets, each corresponding to one section of the join venn diagram:
Cheers,
Ashley W.
Welcome to the Dataiku Community, We are glad to have you with us.
If you are talking about using a visual recipe to do a Join, There are a broad set of joins you can do including traditional SQL type Joins, Fuzzy Joins, and Geo Joins.
In a visual Join Recipe here is the list of types of joins you can do. From your brief description is sounds like you want to do an outer join so you have all records from left data set only , joined data sets, and right dataset only.
There are different visual recipes for windowing, grouping, pivoting and other choices.
If none of these work for you and you can write a bit of SQL you can create a SQL recipe to make whatever kind of join you underlying database infrastructure can support. Not all databases support all kinds of join.
Hope this helps a bit.
Hi @XiShi ,
Great question! You can get unjoined rows (also called an anti-join) using the Join recipe. Here's how:
Similarly, you could also opt for a full outer join like @tgb417 suggested and follow it with a split recipe to create three datasets, each corresponding to one section of the join venn diagram:
Cheers,
Ashley W.
Hi @XiShi ,
We've recently released a new version of Dataiku, 11.3, and you can now add additional output datasets to a Join recipe that contains all the unmatched rows from either a left, right, or inner join.
Cheers,
Ashley