--This SQL script produces a table for the statistic of interest with the defined Get Data Out groups for kaposi sarcomas.
 
--Run on CASREF01 

--Kaposi sarcoma cohort for GDO
--no further splits required, so only contains tumourid, diagnosisyear and age

with tumour_cohort_kaposi 
as 
(
select tumourid
, diagnosisyear
, age
from av2017.at_tumour_england
where site_icd10_o2_3char = 'C46'
and cascade_inci_flag = 1
and diagnosisyear between 2013 and 2017
)


,
statisticofinterest_kaposi as 
(select tumourid, 1 as statisticofinterest_kaposi from tumour_cohort_kaposi),
tumour_cohort_linked_kaposi AS ( SELECT * FROM tumour_cohort_kaposi tc_kaposi
LEFT JOIN statisticofinterest_kaposi ON statisticofinterest_kaposi.tumourid = tc_kaposi.tumourid
)
 
 
 , extractpartition_kaposi as ( 

SELECT '2013' AS "Year", 
SUM(statisticofinterest_kaposi) as statistic_of_interest
FROM tumour_cohort_linked_kaposi
WHERE DIAGNOSISYEAR = '2013'

UNION ALL

SELECT '2014' AS "Year", 
SUM(statisticofinterest_kaposi) as statistic_of_interest
FROM tumour_cohort_linked_kaposi
WHERE DIAGNOSISYEAR = '2014'

UNION ALL

SELECT '2015' AS "Year", 
SUM(statisticofinterest_kaposi) as statistic_of_interest
FROM tumour_cohort_linked_kaposi
WHERE DIAGNOSISYEAR = '2015'

UNION ALL

SELECT '2016' AS "Year", 
SUM(statisticofinterest_kaposi) as statistic_of_interest
FROM tumour_cohort_linked_kaposi
WHERE DIAGNOSISYEAR = '2016'

UNION ALL

SELECT '2017' AS "Year", 
SUM(statisticofinterest_kaposi) as statistic_of_interest
FROM tumour_cohort_linked_kaposi
WHERE DIAGNOSISYEAR = '2017'

UNION ALL

SELECT '2013-2015' AS "Year", 
SUM(statisticofinterest_kaposi) as statistic_of_interest
FROM tumour_cohort_linked_kaposi
WHERE DIAGNOSISYEAR BETWEEN 2013 AND 2015

UNION ALL

SELECT '2014-2016' AS "Year", 
SUM(statisticofinterest_kaposi) as statistic_of_interest
FROM tumour_cohort_linked_kaposi
WHERE DIAGNOSISYEAR BETWEEN 2014 AND 2016

UNION ALL

SELECT '2015-2017' AS "Year", 
SUM(statisticofinterest_kaposi) as statistic_of_interest
FROM tumour_cohort_linked_kaposi
WHERE DIAGNOSISYEAR BETWEEN 2015 AND 2017
 ) ,
  lsoatoregionlookup
as
(
select distinct lsoa11, nhser17nm as Region from nspl_201805
left outer join ANALYSISNCR.CCG18_NHSER17_18_LKP_DATALAKE nr
on nspl_201805.ccg = nr.ccgapr18cd
where substr(lsoa11,1,1) = 'E'
)
, lsoatoalllookup
as
(
select distinct lsoa11, 'All' as Region from nspl_201805
where substr(lsoa11,1,1) = 'E'
)
, geoglookup as 
(select * from lsoatoregionlookup
union
select * from lsoatoalllookup)

/*
reminding myself how dual works:
select rownum from dual;

select rownum  from dual 
 -- LAST YEAR NEEDED
  connect by level <= 3;
*/

,sexlookup
as
(
select 1 as basesex, '1' as sex from dual 
union
select 2 as basesex, '2' as sex from dual 
union
select 1 as basesex, 'All' as sex from dual 
union
select 2 as basesex, 'All' as sex from dual
)

-- year look up
-- This is hacky, there must be a better way? 
-- But it does produce a look up table for 2013-2017
-- and also the three year rolling groups for 2013-2017

,yearlookup
as
(
select n as baseyear
, to_char(n) as year
from
 ( select rownum n  from dual 
 -- LAST YEAR NEEDED
  connect by level <= 2017
 )
 -- FIRST YEAR NEEDED
where n >= 2013
UNION
select n as baseyear
, '2013-2015' as year
from
 ( select rownum n  from dual 
 -- LAST YEAR NEEDED
  connect by level <= 2015
 )
 -- FIRST YEAR NEEDED
where n >= 2013
UNION
select n as baseyear
, '2014-2016' as year
from
 ( select rownum n  from dual 
 -- LAST YEAR NEEDED
  connect by level <= 2016
 )
 -- FIRST YEAR NEEDED
where n >= 2014
UNION
select n as baseyear
, '2015-2017' as year
from
 ( select rownum n  from dual 
 -- LAST YEAR NEEDED
  connect by level <= 2017
 )
 -- FIRST YEAR NEEDED
where n >= 2015
)

-- Making age lookup.

--select * from analysischarlieturner.ages;

-- We don't have a systematic 'we definitely used this age partition'
-- So we should assume any age partition in the age table could be being used.
-- There are currently 30 age partitions.  
-- This is a bit tedious.  

-- You can at least get a list of them:
/*
select COLUMN_NAME from ALL_TAB_COLUMNS
where owner = 'ANALYSISCHARLIETURNER'
and TABLE_NAME = 'AGES'
order by COLUMN_NAME;
*/
-- If I had grip, I would write some R to automate creating this table
-- Instead I did it in Excel
, agelookup
as
(
select distinct floor(AGES.AGE/5)+1 as baseage, 'All' as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_13_25_40_60_80 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_20_30_50_70 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_20_40 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_20_50_70 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_20_50_70_80 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_20_TO_40 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_20_TO_90 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_25 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_25_40_60 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_25_40_60_80 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_25_40_TO_80 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_25_50 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_30 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_30_50 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_30_50_70 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_30_50_70_80 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_30_TO_50 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_30_TO_70 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_35 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_40 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_40_TO_70 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_5_THEN_10 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_50 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_50_70 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_50_70_80 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_50_TO_70 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_50_TO_80 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_5_10_20_30_50_70 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_60 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_60_70 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_60_70_75_80 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_60_TO_80 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_70 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_70_75_80 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_70_80 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, SPLIT_80 as age from analysischarlieturner.ages
UNION
select distinct floor(AGES.AGE/5)+1 as baseage, TENYEARAGE as age from analysischarlieturner.ages)

-- So, I have a region lookup, a sex lookup, a year lookup, and an age lookup.

-- We have no joins for the Welsh LSOAs and for the years before 2013. 
-- But that is what we wanted
-- All ages and sexs have joined to something.  


, population_table as
(
select g.region, s.sex, y.year, a.age, sum(popcount) as popcount from ons2017.populations_normalised POPS
left outer join geoglookup g 
on pops.lsoa11 = g.lsoa11
left outer join sexlookup s
on pops.sex = s.basesex
left outer join yearlookup y
on pops.year = y.baseyear 
left outer join agelookup a
on pops.quinaryagegroupint = a.baseage
group by g.region, s.sex, y.year, a.age
)
, 
kaposi as(
SELECT
'Kaposi Sarcoma' AS "Cancer Site",
"Year" AS "Year",
'All' AS "Tumour Type",
'All' AS "Tumour Type 2",
'All' AS "Tumour Type 3",
'All' AS "Basis of Diagnosis",
'Not applicable' AS "Stage",
'Not split by stage detail' AS "Stage detail",
'All ages' AS "Age",
'All England' AS "Region",
'Persons' AS "Sex",
'Not applicable' AS "Grade",
'Not applicable' AS "Nodal Status",
statistic_of_interest as "statistic_of_interest",
pops.popcount  as "Population"
from extractpartition_kaposi ep
left outer join population_table pops
on ep."Year" = pops.year
and pops.age = 'All'
where pops.region ='All'
and pops.sex='All'
)
 select * from kaposi ;