问题
I created a coldfusion script allowing to get an object "unit" from an array.
My DB has a view about units. For a same unit key "ORG_ID", it can exist several rows (with a difference on a field "origin"). the field "origin" can be "current", "history" or "different".
+---------+---------+------------+------------+----------|
| ORG_ID | TITLE | VALID_FROM | VALID_TO | ORIGIN |
+---------+---------+------------+------------+----------|
| 1234 | A.1 | 01/03/2016 | 31/12/9999 | other |
| 1234 | A.1 | 01/03/2016 | 31/12/3333 | current |
| 1234 | A.1 | 01/03/2016 | 31/12/9999 | history |
| 5420 | A.2 | 01/01/2014 | 31/12/3333 | other |
| 9876 | A.3 | 01/03/2016 | 31/12/3333 | current |
| 9876 | B.3 | 01/03/2016 | 31/12/9999 | history |
| 5527 | A.1 | 01/03/2016 | 31/12/2199 | current |
| 5527 | D.2 | 01/01/2010 | 31/12/2015 | history |
| 5527 | A.1 | 01/01/2016 | 31/12/2199 | history |
| 6699 | E.5 | 01/01/2016 | 31/12/2017 | history |
| 6699 | A.4 | 01/01/2017 | 31/12/2018 | history |
+---------+---------+------------+------------+----------|
In this case for instance here the result that I would like to get:
+---------+---------+------------+------------+----------|--------------|
| ORG_ID | TITLE | VALID_FROM | VALID_TO | ORIGIN | CORRECT_VERS |
+---------+---------+------------+------------+----------|--------------|
| 1234 | A.1 | 01/03/2016 | 31/12/9999 | other | 0 |
| 1234 | A.1 | 01/03/2016 | 31/12/3333 | current | 1 |
| 1234 | A.1 | 01/03/2016 | 31/12/9999 | history | 0 |
| 5420 | A.2 | 01/01/2014 | 31/12/3333 | other | 1 |
| 9876 | A.3 | 01/03/2016 | 31/12/3333 | current | 1 |
| 9876 | B.3 | 01/03/2016 | 31/12/9999 | history | 0 |
| 5527 | A.1 | 01/03/2016 | 31/12/2199 | current | 1 |
| 5527 | D.2 | 01/01/2010 | 31/12/2015 | history | 0 |
| 5527 | A.1 | 01/01/2016 | 31/12/2199 | history | 0 |
| 6699 | E.5 | 01/01/2016 | 31/12/2017 | history | 0 |
| 6699 | A.4 | 01/01/2017 | 31/12/2018 | history | 0 |
+---------+---------+------------+------------+----------+--------------|
My Coldfusion script: dataUnitArray contains the list of units in an array
<cftry>
<cfset hist = 0/>
<cfset unit = structNew() />
<cfloop index="i" from="1" to="#ArrayLen(dataUnitArray)#">
<cfif #dataUnitArray[i].ORIGIN# EQ "current">
<!--- Unit is current --->
<cfscript>
unit.ORG_ID = #dataUnitArray[i].ORG_ID#;
unit.TITLE = #dataUnitArray[i].TITLE#;
unit.UNIT_VALID_FROM = #dateFormat(dataUnitArray[i].UNIT_VALID_FROM, 'DD/MM/YYYY')#;
unit.UNIT_VALID_TO = #dateFormat(dataUnitArray[i].UNIT_VALID_TO, 'DD/MM/YYYY')#;
unit.ORIGIN = #dataUnitArray[i].ORIGIN#;
return unit;
</cfscript>
<cfelse>
<cfif #dataUnitArray[i].ORIGIN# EQ "history">
<!--- Unit is history --->
<cfscript>
unit.ORG_ID = #dataUnitArray[i].ORG_ID#;
unit.TITLE = #dataUnitArray[i].TITLE#;
unit.UNIT_VALID_FROM = #dateFormat(dataUnitArray[i].UNIT_VALID_FROM, 'DD/MM/YYYY')#;
unit.UNIT_VALID_TO = #dateFormat(dataUnitArray[i].UNIT_VALID_TO, 'DD/MM/YYYY')#;
unit.ORIGIN = #dataUnitArray[i].ORIGIN#;
</cfscript>
<cfset hist++ >
<cfelse>
<!--- Unit is different (other) --->
<cfif hist EQ 0>
<cfscript>
unit.ORG_ID = #dataUnitArray[i].ORG_ID#;
unit.TITLE = #dataUnitArray[i].TITLE#;
unit.UNIT_VALID_FROM = #dateFormat(dataUnitArray[i].UNIT_VALID_FROM, 'DD/MM/YYYY')#;
unit.UNIT_VALID_TO = #dateFormat(dataUnitArray[i].UNIT_VALID_TO, 'DD/MM/YYYY')#;
unit.ORIGIN = #dataUnitArray[i].ORIGIN#;
</cfscript>
</cfif>
</cfif>
</cfif>
</cfloop>
<cfscript>
return unit;
</cfscript>
<cfcatch type="any">
<cfscript>
.....
</cfscript>
</cfcatch>
</cftry>
My script is correctly working. But I have loading time problem when I used it on a lot of data. That's why I would like to do that directly in ORACLE, with CASE...WHEN as:
CASE
when ORIGIN = 'current' THEN 1
WHEN ORIGIN = 'history' THEN
CASE hist = 0 THEN ....
END
ELSE
0
END AS "IS_CORRECT_VERSION"
I would like to add a new column "CORRECT_VERSION" ( value 0 or 1 when the version is correct) in a view in order to retrieve the correct unit version.
But I don't know how to do that, could you please help me with that?
Thanks in advance for your help.
Seb
回答1:
I do not know ColdFusion, but I think I understood the logic. Priority is current > history > different. It is not clear which row is correct when there are two current rows or only different rows, so I mark row with minimum valid_from in such case. If you don't care You can omit this parameter (remove unit_valid_from from row_number's order by clause):
select units.*,
case when 1 =
row_number() over (
partition by org_id
order by case origin when 'current' then 1 when 'history' then 2 else 3 end,
unit_valid_from ) then 1 else 0 end as is_correct_version
from units
dbfiddle demo
回答2:
I assume that there are no two rows having the same origin for the same org_id and even if duplicates exist, there must be only one record has VALID_FROM < SYSDATE < VALID_TO.
Following is the solution based on sample data. Please change the logic if required.
SELECT
UNITS.*,
CASE
WHEN DENSE_RANK() OVER(
PARTITION BY ORG_ID
ORDER BY
CASE ORIGIN
WHEN 'current' THEN 1
WHEN 'history' THEN 2
ELSE 3
END
) = 1
AND TRUNC(SYSDATE) BETWEEN VALID_FROM AND VALID_TO THEN 1
ELSE 0
END AS IS_CORRECT_VERSION
FROM
UNITS;
Cheers!!
来源:https://stackoverflow.com/questions/57460093/convert-coldfusion-script-in-oracle-view-with-case-and-iteration