Join same column from multiple tables - sql

Question

Below is my current code. I'm not sure what the best way is to amend this to give me the results I need.
SELECT
T1.SC,
T1.AN,
T1.DOFS_DATE,
T2.M_ID,
T3.OPDT,
T4.MARKER,
T5.E_DTE,
T5.E_TME,
T5.E_PST_DTE,
T5.E_AMT,
T5.E_NAR_O,
T5.E_NAR_T
FROM E_Base.AR_MyTable T1
LEFT JOIN E_Base.Translation T2
ON T1.SC = T2.SC
AND T1.AN = T2.AN
LEFT JOIN E_Base.BA T3
ON T2.M_ID = T3.M_ID
LEFT JOIN E_Base.APF T4
ON T3.M_ID = T4.M_ ID
AND MARKER = 54
LEFT JOIN U_DB.TEH_201804 T5
ON T2.M_ID = T5.M_ID
AND T1.DOFS_DATE = T5.E_PST_DTE
QUALIFY ROW_NUMBER() OVER (PARTITION BY T2.M_ID ORDER BY T2.ID_END_DATE DESC, T3.E_END_DATE DESC) = 1
The above code works. However, it is the final left join on T5 where I need help.
In T1 each M_ID has assigned it's own DOFS_DATE that could be any date within the year and I want the data from T5 U_DB.TEH_201804 for the matching date. However, 5 U_DB.TEH_201804 relates to only April 2018. There are 12 tables with the same database (201804, 201805, 201806 etc) that all have the exact same columns but relate to a different month within the year.
Ideally, I want to left join the columns from T5 once but search all 12 tables within the database to bring back the data where the dates correspond.
I was thinking UNION but am unsure how to work this in.
Any help would be greatly appreciated!
Thanks

Answer 1

You could change you code related to table t5 wuth a left join on a subquery that select the union all for all the bale you need ...... (i have named the subquery TT)
SELECT
T1.SC,
T1.AN,
T1.DOFS_DATE,
T2.M_ID,
T3.OPDT,
T4.MARKER,
TT.E_DTE,
TT.E_TME,
TT.E_PST_DTE,
TT.E_AMT,
TT.E_NAR_O,
TT.E_NAR_T
FROM E_Base.AR_MyTable T1
LEFT JOIN E_Base.Translation T2
ON T1.SC = T2.SC
AND T1.AN = T2.AN
LEFT JOIN E_Base.BA T3
ON T2.M_ID = T3.M_ID
LEFT JOIN E_Base.APF T4
ON T3.M_ID = T4.M_ ID
AND MARKER = 54
LEFT JOIN (
select *
FROM U_DB.TEH_201804
UNION ALL
select *
FROM U_DB.TEH_201805
UNION ALL
select *
FROM U_DB.TEH_201806
UNION ALL
select *
FROM U_DB.TEH_201807
UNION ALL
.....
) TT ON T2.M_ID = TT.M_ID
AND T1.DOFS_DATE = TT.E_PST_DTE
QUALIFY ROW_NUMBER() OVER (PARTITION BY T2.M_ID ORDER BY T2.ID_END_DATE DESC, T3.E_END_DATE DESC) = 1

Answer 2

It's hard to tell without additional details like explain and QueryLog step data.
Based on #scaisEdge answer:
You can try to move the first two joins into a Derived Table to apply the ROW_NUMBER early (possible because you do Outer Joins only):
SELECT
dt.*,
T4.MARKER,
TT.E_DTE,
TT.E_TME,
TT.E_PST_DTE,
TT.E_AMT,
TT.E_NAR_O,
TT.E_NAR_T
FROM
(
SELECT
T1.SC,
T1.AN,
T1.DOFS_DATE,
T2.M_ID,
T3.OPDT
FROM E_Base.AR_MyTable T1
LEFT JOIN E_Base.Translation T2
ON T1.SC = T2.SC
AND T1.AN = T2.AN
LEFT JOIN E_Base.BA T3
ON T2.M_ID = T3.M_ID
QUALIFY Row_Number()
Over (PARTITION BY T2.M_ID
ORDER BY T2.ID_END_DATE DESC, T3.E_END_DATE DESC) = 1
) AS dt
LEFT JOIN E_Base.APF T4
ON dt.M_ID = T4.M_ID
AND MARKER = 54
LEFT JOIN
(
SELECT *
FROM U_DB.TEH_201804
UNION ALL
SELECT *
FROM U_DB.TEH_201805
UNION ALL
SELECT *
FROM U_DB.TEH_201806
UNION ALL
SELECT *
FROM U_DB.TEH_201807
UNION ALL
.....
) TT
ON dt.M_ID = TT.M_ID
AND dt.DOFS_DATE = TT.E_PST_DTE
It might also help the optimizer to provide additional info about the data ranges. Those tables should have CHECK-constraints to tell the optimizer that they contain only data from a single month, if they don't exist try adding a WHERE-condition to each Select, e.g. WHERE E_PST_DTE BETWEEN DATE '2018-04-01' AND DATE '2018-04-30'.
Of course, always check Explain if the plan actually changes...