How to merge 2 tables into 1 row, with multiple entries from second table in Oracle SQL - sql

Question

newbee question on (Oracle) SQL.
I'd like this table :
ash_id ash_contact_name ash_contact_telefoonnummber
15313 Name1 022457852114
15313 Name2 122457852114
15313 Name3 222457852114
15313 Name4 322457852114
15313 Name5 422457852114
To Look like this in 1 row :
15313 Name1 022457852114 Name2 122457852114 Name3 222457852114 Name4 322457852114 Name5 422457852114
So I get only 1x the id from first table and multiple coloms with the name with My code now looks like this :
select ash.ash_id ,
con.ash_contact_name, con.ash_contact_telefoonnummer
from D00ASH01.ash_admin_stakeholder ash,ash_contacts con
where con.ash_id = ash.ash_id and con.ash_id = 15313
order by ash.ash_id
The eventual code while not include "con.ash_id = 15313" as I will need to have all the entries. The end result will include more fields from the first table, so I can not just use the second table alone. For now, I want to start to build it up simple.
I tried to make it work with a join but did not made it.
All suggestions welcome,
thanks

Answer 1

Check this, it might help you. If it is not then let me know.
SELECT ash.ash_id , con.ash_contact_name, con.ash_contact_telefoonnummer FROM D00ASH01.ash_admin_stakeholder as ash INNER JOIN ash_contacts as con ON con.ash_id = ash.ash_id WHERE con.ash_id = 15313 ORDER BY ash.ash_id;

Answer 2

SELECT * FROM ash INNER JOIN con USING( ash_id ) Where con_id = 15313 ORDER BY ash_id ;

Answer 3

I'did get the solution for this :
select
c.ash_id,
c.ash_naam_kbo_NL ,
c.ash_naam_kbo_FR ,
c.ash_naam_kbo_DE,
max(Case when rn = 1 then c.ASH_CONTACT_NAME else null end) as name1,
max(Case when rn = 1 then c.ASH_CONTACT_GSMNUMMER else null end) as gsm1,
max(Case when rn = 1 then c.ASH_CONTACT_FAXNUMMER else null end) as fax1,
max(Case when rn = 1 then c.ASH_CONTACT_EMAILADRES else null end) as email1,
max(Case when rn = 2 then c.ASH_CONTACT_NAME else null end) as name2,
max(Case when rn = 2 then c.ASH_CONTACT_GSMNUMMER else null end) as gsm2,
max(Case when rn = 2 then c.ASH_CONTACT_FAXNUMMER else null end) as fax2,
max(Case when rn = 2 then c.ASH_CONTACT_EMAILADRES else null end) as email2
from (
select
table_ash.ash_id,
table_ash.ash_naam_kbo_NL ,
table_ash.ash_naam_kbo_FR ,
table_ash.ash_naam_kbo_DE,
table_contacts.ASH_CONTACT_NAME,
table_contacts.ASH_CONTACT_GSMNUMMER,
table_contacts.ASH_CONTACT_FAXNUMMER,
table_contacts.ASH_CONTACT_EMAILADRES,
ROW_NUMBER () over (partition by table_ash.ash_id order by table_ash.ash_id,
table_ash.ash_naam_kbo_NL) rn
from
ASH_ADMIN_STAKEHOLDER table_ash,
ash_contacts table_contacts
where
table_ash.ash_id = table_contacts.ash_id) c
group by
c.ash_id,
c.ash_naam_kbo_NL ,
c.ash_naam_kbo_FR ,
c.ash_naam_kbo_DE;
So, there's a "select" from a "select". The trick is to generate a rownnumber and use it like a index. For as many as rownumbers as needed, include a max function in the query. This code is the base of answer I needed.

Answer 4

SQL code written in Oracle:
WITH CTE AS(
SELECT
            UP.CLASS,
            UP.NS || UP.RN AS NSR,
            UP.VAL
FROM
            (
            SELECT
                        ROW_NUMBER ()
         OVER (
           PARTITION BY S.CLASS
            ORDER BY
                        S.CLASS) RN,
                        S.*
            FROM
                        STAKEHOLDER S
            ORDER BY
                        CLASS,
                        SID) SS
UNPIVOT (VAL FOR NS IN (NAME, SID)) UP
)
SELECT
            *
FROM
            CTE
PIVOT(MAX(VAL) FOR NSR IN ('NAME1' AS NAME1,
            'SID1' AS SID1,
            'NAME2' AS NAME2,
            'SID2' AS SID2,
            'NAME3' AS NAME3,
            'SID3' AS SID3))
This is not difficult if we handle it with our natural way of thinking. After grouping the table by CLASS, we convert NAME and SID columns into rows and create names commanding values to be converted to columns. Format of names is the original column name + number of subgroups, like NAME1, SID1, NAME2, SID2，… for group 1 and NAME1, SID1, … for group2. Then we concatenate groups and transpose row to columns. The problem is SQL does not support dynamic row-to-column/column-to-row transposition. When the number of columns is small and columns are fixed, the language can mange to do the transpositions. As the number of columns increases, the scenario becomes more and more awkward. Enumerating all columns to be converted is complicated and SQL code becomes bloated. If columns are dynamic, SQL needs to turn to complex and roundabout ways to handle them.
Yet, it is really easy to code the transposition task with the open-source esProc SPL:
| |A|
|1|=connect("ORACLE")|
|2|=A1.query#x("SELECT \* FROM STAKEHOLDER ORDER BY CLASS,SID")|
|3|=A2.fname().m(2:)|
|4|=A2.group#o(CLASS)|
|5|=A4.conj(\~.news(A3;CLASS,A3(#)/A4.\~.#:COL,\~:VAL))|