Is it possible to join 2 tables that have no common table?

Hi Lydia,

If there is an indirect relation being [Table A]N <->1[Table B]1 <->N[Table C] and you want to combine Table A and C, BIND should work.

If there is not such a connection (indirectly), only a rearrangement of the data model would be a solution here, to match it with the documentation (link).

If you really want to keep the current set-up and also want to add more information, it is sometimes usefull to (partially) duplicate a table and place it twice in the data model.

Hi Jan-peter!

Thank you for the feedback. We did try to change [KNA1]N -> 1[AR_KNVV_KNVP] but did not work. Do you have some guide on how we can duplicate the table (duplicating the AR_KNVV table here?) and place it twice in the data model?

Hi Lydia,

Using this script you can copy a table and name it whatever you want as long as the name is unique (i.e. AR_KNVV_DUPLICATE). Then you can add it to the existing data model.

DROP TABLE IF EXISTS "<table_duplicate>";

CREATE TABLE "<table_duplicate>" AS SELECT * FROM "<table_original>";

Note that if you use name mapping in your data model, the duplicate table is not recognized, due to a different naming. You could fix this by adding the entries for this table manually.

Hi @1460047367 

Yes in such cases BIND functionality do work.

I have faced a similar situation where in there was no direct link between table VBKD and T171T table and then I have used BIND function keeping case table as target table .

below is the syntax.

BIND("VBAP","VBKD"."BZIRK"||' - '||"T171T"."BZTXT")

You can try this out, hope it works.

Hi @janpeter.van.d. I have a doubt regarding the use of BIND in the case of this relationship [Table A]N <->1[Table B]1 <->N[Table C]. How can bind be use din this case? I have read the documentation and it states BIND is appropriate for 1:N:1 relationships.

Hi @laura.santo12,

You're right. I've become more familiar with the bind function since the time of this post, and would say that maybe a normal PU-function (Pull Up Aggregation (celonis.com)) would maybe be sufficient here.

Kind regards,

Jan-peter

hI @ROHIT KUMAR 

Could you please further explain which was the N:1 1:N relationship here (meaning which tables were 1 and N). Also, I tried this with DATE fields and it says that CONCAT do not support date fields.

Many thanks