Dedicated SQL file

sql/queries.pgsql

@@ -0,0 +1,99 @@
+-- Create "tall" table for april 2017 (from Jason) 
+CREATE TABLE public.coup_tall_april AS
+SELECT  a.icp_id
+     , a.read_date
+     , c.period
+     , sum(c.read_kwh) as kwh_tot
+     , sum(case when a.content_code = 'UN' then c.read_kwh else 0 end) as kwh_un
+     , sum(case when a.content_code in ('CN','EG') then c.read_kwh else 0 end) as kwh_cn
+FROM    coup_prd.coupdatamaster a,
+	unnest(a.read_array) WITH ORDINALITY c(read_kwh, period)
+WHERE   a.read_date >= to_date('01/04/2017','dd/mm/yyyy')
+ and   a.read_date <  to_date('01/05/2017','dd/mm/yyyy')
+ and   a.content_code  ~ ('UN|CN|EG')
+GROUP BY 1, 2, 3
+ORDER BY 1, 2, 3;
+
+-- "Head" of the data
+SELECT * FROM public.coup_tall_april limit 10;
+
+-- Count distinct ICPs for each date
+SELECT read_date, COUNT(DISTINCT icp_id) AS d_icp 
+FROM public.coup_tall_april 
+GROUP BY read_date;
+
+-- Calculate mininmums, maximums, averages for each date
+SELECT read_date, MIN(kwh_tot), AVG(kwh_tot), MAX(kwh_tot) 
+FROM public.coup_tall_april 
+GROUP BY read_date;
+
+-- Find negative values
+SELECT * FROM public.coup_tall_april WHERE kwh_tot < 0 OR kwh_un < 0 OR kwh_cn < 0;
+
+-- Count number of ICPs that end in "17"
+SELECT COUNT (DISTINCT icp_id) FROM public.coup_tall_april WHERE icp_id LIKE '%17';
+
+-- List first 10 ICPs that end with number 17, sorted
+SELECT DISTINCT icp_id FROM public.coup_tall_april WHERE icp_id LIKE '%17' ORDER BY icp_id LIMIT 10;
+
+-- Verify that all ICP IDs are of length 7
+SELECT MIN(CHAR_LENGTH(icp_id)) as minl, MAX(CHAR_LENGTH(icp_id)) as maxl FROM public.coup_tall_april;
+
+-- String conversion
+SELECT DISTINCT CAST(SUBSTRING(icp_id FROM 2 FOR 6) AS int) AS icp_val, icp_id FROM public.coup_tall_april 
+
+-- Count nulls in cast (none)
+SELECT COUNT(*) - COUNT(CAST(SUBSTRING(icp_id FROM 2 FOR 6) AS int)) FROM public.coup_tall_april;
+
+-- Verify that there are only "period" values from 1 to 48
+SELECT DISTINCT period FROM public.coup_tall_april ORDER BY period;
+
+-- Look at the 2nd specifically (DST ends, 2017)
+SELECT * FROM public.coup_tall_april 
+WHERE read_date = to_date('02/04/2017', 'dd/mm/yy') 
+ORDER BY icp_id, period 
+LIMIT 60;
+
+-- September interlude STARTS
+-- Create "tall" table for Septeber 24 2017 (mod from Jason, above) to look at DST starting
+CREATE TABLE public.coup_dst_2017 AS
+SELECT  a.icp_id
+     , a.read_date
+     , c.period
+     , sum(c.read_kwh) as kwh_tot
+     , sum(case when a.content_code = 'UN' then c.read_kwh else 0 end) as kwh_un
+     , sum(case when a.content_code in ('CN','EG') then c.read_kwh else 0 end) as kwh_cn
+FROM    coup_prd.coupdatamaster a,
+	unnest(a.read_array) WITH ORDINALITY c(read_kwh, period)
+WHERE   a.read_date = to_date('24/09/2017','dd/mm/yyyy')
+ and   a.content_code  ~ ('UN|CN|EG')
+GROUP BY 1, 2, 3
+ORDER BY 1, 2, 3;
+
+-- Look at September 24 2017 table
+SELECT * FROM public.coup_dst_2017 
+ORDER BY icp_id, period 
+LIMIT 60;
+
+-- Verify that there are only "period" values from 1 to 48
+SELECT COUNT(DISTINCT period) FROM public.coup_dst_2017;
+-- September interlude ENDS
+
+-- Timestamp conversion
+SELECT DISTINCT DATE_PART('year', read_date), DATE_PART('month', read_date), 
+    DATE_PART('day', read_date) FROM public.coup_tall_april ORDER BY 1, 2, 3;
+
+-- Creating "time" values
+SELECT *, (CONCAT(q.h, ':', q.m, ':00'))::time AS t FROM
+(
+    SELECT DISTINCT period, period / 2 as h, period % 2 * 30 AS m
+    FROM public.coup_tall_april
+) AS q ORDER BY period;
+
+-- Creating timestamp column
+SELECT *, read_date + CONCAT(period / 2, ':', period % 2 * 30, ':00')::time AS read_time
+FROM public.coup_tall_april ORDER BY icp_id, read_time limit 50;
+
+-- Possible subset for further testing
+SELECT *, read_date + CONCAT(period / 2, ':', period % 2 * 30, ':00')::time AS read_time
+FROM public.coup_tall_april WHERE icp_id LIKE '%1117' ORDER BY icp_id, read_time;