I'm trying to return some results spread over a rolling 12 month period eg:
MONTH IN OUT
January 210 191
February 200 111
March 132 141
April 112 141
May 191 188
etc...
How do I spread the results over a date range, populating the first column with the month name?
IN MSSQL it would be something like:
SELECT COUNT(problem.problem_type = 'IN') AS IN,
COUNT(problem.problem_type = 'OUT') AS OUT,
DATEPART(year, DateTime) as Year,
DATEPART(month, DateTime) as Month
FROM problem
WHERE (DateTime >= dbo.FormatDateTime('2010-01-01'))
AND
(DateTime < dbo.FormatDateTime('2010-01-31'))
GROUP BY DATEPART(year, DateTime),
DATEPART(month, DateTime);
But this is against an Oracle database so DATEPART and DateTime are not available.
My Problem table is roughly:
problem_ID Problem_type IN_Date OUT_Date
1 IN 2010-01-23 16:34:29.0 2010-02-29 13:06:28.0
2 IN 2010-01-27 12:34:29.0 2010-01-29 12:01:28.0
3 OUT 2010-02-13 13:24:29.0 2010-09-29 15:04:28.0
4 OUT 2010-02-15 16:31:29.0 2010-07-29 11:03:28.0
Use:
SELECT SUM(CASE WHEN p.problem_type = 'IN' THEN 1 ELSE 0 END) AS IN,
SUM(CASE WHEN p.problem_type = 'OUT' THEN 1 ELSE 0 END) AS OUT,
TO_CHAR(datetime, 'YYYY') AS year,
TO_CHAR(datetime, 'MM') AS month
FROM PROBLEM p
WHERE p.DateTime >= TO_DATE('2010-01-01', 'YYYY-MM-DD')
AND p.DateTime < TO_DATE('2010-01-31', 'YYYY-MM-DD')
GROUP BY TO_CHAR(datetime, 'YYYY'), TO_CHAR(datetime, 'MM')
You could also use:
SELECT SUM(CASE WHEN p.problem_type = 'IN' THEN 1 ELSE 0 END) AS IN,
SUM(CASE WHEN p.problem_type = 'OUT' THEN 1 ELSE 0 END) AS OUT,
TO_CHAR(datetime, 'MM-YYYY') AS mon_year
FROM PROBLEM p
WHERE p.DateTime >= TO_DATE('2010-01-01', 'YYYY-MM-DD')
AND p.DateTime < TO_DATE('2010-01-31', 'YYYY-MM-DD')
GROUP BY TO_CHAR(datetime, 'MM-YYYY')
Reference:
TO_CHAR
TO_DATE
You probably want something like
SELECT SUM( (CASE WHEN problem_type = 'IN' THEN 1 ELSE 0 END) ) in,
SUM( (CASE WHEN problem_type = 'OUT' THEN 1 ELSE 0 END) ) out,
EXTRACT( year FROM DateTime ) year,
EXTRACT( month FROM DateTime ) month
FROM problem
WHERE DateTime >= date '2010-01-01'
AND DateTime < date '2010-01-31'
GROUP BY EXTRACT( year FROM DateTime ),
EXTRACT( month FROM DateTime )
Related
I'm trying to return a couple of 'average if' columns using the following:
select
date,
avg(case when hour >= 23 or hour <= 6) then (select price) else null end) as price1,
avg(case when (hour >= 16 and hour <= 18) then (select price) else null end) as price2
from
xxxxxxxxx
where
date <= '2019-12-31' and
date >= '2018-12-01'
group by
date
order by
date
it works when I use each avg(case when) individually but when I use them both I get the error
Invalid operation: This type of correlated subquery pattern is not supported due to internal error
Why the select in the select?
select date,
avg(case when hour >= 23 or hour <= 6 then price end) as price1,
avg(case when hour >= 16 and hour <= 18 then price end) as price2
from xxxxxxxxx
where date <= '2019-12-31' and
date >= '2018-12-01'
group by date
order by date;
The else null is also redundant.
I need count in range two date，this sql is work，bug not better,can you help me?
select dmc.doctor_id,
(
select count(*)
from hele_dct_member_config dmc
WHERE (EXTRACT(YEAR FROM dmc.start_time) = 2016 OR EXTRACT(YEAR FROM dmc.end_time) = 2016) AND dmc.status=1
AND TO_DATE('2016-01-31', 'yyyy-mm-dd') BETWEEN start_time AND end_time
) Jan,
(
select count(*)
from hele_dct_member_config dmc
WHERE (EXTRACT(YEAR FROM dmc.start_time) = 2016 OR EXTRACT(YEAR FROM dmc.end_time) = 2016) AND dmc.status=1
AND TO_DATE('2016-02-28', 'yyyy-mm-dd') BETWEEN start_time AND end_time
) Feb,
.
.
.
from hele_dct_member_config dmc
enter code here
WHERE (EXTRACT(YEAR FROM dmc.start_time) = 2016 OR EXTRACT(YEAR FROM dmc.end_time) = 2016) AND dmc.status=1
grouy by dmc.doctor_id
I need count in range two date，this sql is work，bug not better,can you help me?
Use conditional aggregation:
select dmc.doctor_id,
sum(case when date '2016-01-31' BETWEEN start_time AND end_time then 1 else 0
end) as Jan,
sum(case when date '2016-02-31' BETWEEN start_time AND end_time then 1 else 0
end) as Feb,
.
.
.
from hele_dct_member_config dmc
where (extract(year from dmc.start_time) = 2016 or
extract(year from dmc.end_time) = 2016) AND
dmc.status = 1
group by dmc.doctor_id;
if i want get quarter count ? this is one way
SUM(case when date '2016-01-31' BETWEEN start_time AND end_time then 1 else 0 end) +
SUM(case when date '2016-02-28' BETWEEN start_time AND end_time then 1 else 0 end) +
SUM(case when date '2016-03-31' BETWEEN start_time AND end_time then 1 else 0 end) one
I have a table like:
ID TIMEVALUE
----- -------------
1 06.07.15 06:43:01,000000000
2 06.07.15 12:17:01,000000000
3 06.07.15 18:21:01,000000000
4 06.07.15 23:56:01,000000000
5 07.07.15 04:11:01,000000000
6 07.07.15 10:47:01,000000000
7 07.07.15 12:32:01,000000000
8 07.07.15 14:47:01,000000000
and I want to group this data by special times.
My current query looks like this:
SELECT TO_CHAR(TIMEVALUE, 'YYYY\MM\DD'), COUNT(ID),
SUM(CASE WHEN TO_CHAR(TIMEVALUE, 'HH24MI') <=700 THEN 1 ELSE 0 END) as morning,
SUM(CASE WHEN TO_CHAR(TIMEVALUE, 'HH24MI') >700 AND TO_CHAR(TIMEVALUE, 'HH24MI') <1400 THEN 1 ELSE 0 END) as daytime,
SUM(CASE WHEN TO_CHAR(TIMEVALUE, 'HH24MI') >=1400 THEN 1 ELSE 0 END) as evening FROM Table
WHERE TIMEVALUE >= to_timestamp('05.07.2015','DD.MM.YYYY')
GROUP BY TO_CHAR(TIMEVALUE, 'YYYY\MM\DD')
and I am getting this output
day overall morning daytime evening
----- ---------
2015\07\05 454 0 0 454
2015\07\06 599 113 250 236
2015\07\07 404 139 265 0
so that is fine grouping on the same day (0-7 o'clock, 7-14 o'clock and 14-24 o'clock)
But my question now is:
How can I group over midnight?
For example count from 6-14 , 14-23 and 23-6 o'clock on next day.
I hope you understand my question. You are welcome to even improve my upper query if there is a better solution.
EDIT: It is tested now: SQL Fiddle
The key is simply to adjust the group by so that anything before 6am gets grouped with the previous day. After that, the counts are pretty straight-forward.
SELECT TO_CHAR(CASE WHEN EXTRACT(HOUR FROM timevalue) < 6
THEN timevalue - 1
ELSE timevalue
END, 'YYYY\MM\DD') AS day,
COUNT(*) AS overall,
SUM(CASE WHEN EXTRACT(HOUR FROM timevalue) >= 6 AND EXTRACT(HOUR FROM timevalue) < 14
THEN 1 ELSE 0 END) AS morning,
SUM(CASE WHEN EXTRACT(HOUR FROM timevalue) >= 14 AND EXTRACT(HOUR FROM timevalue) < 23
THEN 1 ELSE 0 END) AS daytime,
SUM(CASE WHEN EXTRACT(HOUR FROM timevalue) < 6 OR EXTRACT(HOUR FROM timevalue) >= 23
THEN 1 ELSE 0 END) AS evening
FROM my_table
WHERE timevalue >= TO_TIMESTAMP('05.07.2015','DD.MM.YYYY')
GROUP BY TO_CHAR(CASE WHEN EXTRACT(HOUR FROM timevalue) < 6
THEN timevalue - 1
ELSE timevalue
END, 'YYYY\MM\DD');
Substract 1 day from timevalue for times lower than '06:00' at first and then:
SQLFiddle demo
select TO_CHAR(day, 'YYYY\MM\DD') day, COUNT(ID) cnt,
SUM(case when '23' < tvh or tvh <= '06' THEN 1 ELSE 0 END) as midnight,
SUM(case when '06' < tvh and tvh <= '14' THEN 1 ELSE 0 END) as daytime,
SUM(case when '14' < tvh and tvh <= '23' THEN 1 ELSE 0 END) as evening
FROM (
select id, to_char(TIMEVALUE, 'HH24') tvh,
trunc(case when (to_char(timevalue, 'hh24') <= '06')
then timevalue - interval '1' day
else timevalue end) day
from t1
)
GROUP BY day
Maybe you can do it like this (with some reformatting or PIVOT):
WITH spans AS
(SELECT TIMESTAMP '2015-01-01 00:00:00' + LEVEL * INTERVAL '1' HOUR AS start_time
FROM dual
CONNECT BY TIMESTAMP '2015-01-01 00:00:00' + LEVEL * INTERVAL '1' HOUR < LOCALTIMESTAMP),
t AS
(SELECT start_time, lead(start_time, 1) OVER (ORDER BY start_time) AS end_time, ROWNUM AS N
FROM spans
WHERE EXTRACT(HOUR FROM start_time) IN (6,14,23))
SELECT N, start_time, end_time, COUNT(*) AS ID_COUNT,
DECODE(EXTRACT(HOUR FROM start_time), 6,'morning', 14,'daytime', 23,'evening') AS daytime
FROM t
JOIN YOUR_TABLE WHERE TIMEVALUE BETWEEN start_time AND end_time
GROUP BY N;
Of course, the initial time value ('2015-01-01 00:00:00' in my example) has to be lower than the least date in your table.
I have a membership table with the following columns:
Member_number | StartDate | EndDate
XYZ | 01-Jan-2002 | 01-March-2002
ABC | 01-Feb-2002 | 01-March-2002
Basically, I want to show how many members were present in specific month. My problem is I don't know how to break this time span into months. How can I see this result?
Month | NumberOfMembers
Jan | 1
Feb | 2
March | 2
Given a members table that looks something like this:
create table dbo.members
(
member_number int not null primary key ,
start_date datetime not null ,
end_date datetime not null ,
)
And a table-valued function that generates sequences of consecutive integers, like this:
create function dbo.IntegerRange ( #from int , #thru int )
returns #sequence table
(
value int not null primary key clustered
)
as
begin
declare #increment int = case when #from > #thru then -1 else 1 end ;
with sequence(value) as
(
select value = #from
union all
select value + #increment
from sequence
where value < #thru
)
insert #sequence
select value
from sequence
order by value
option ( MAXRECURSION 0 )
return
end
A query like this should give you what you want:
select [year] = period.yyyy ,
[month] = case period.mm ,
when 1 then 'Jan'
when 2 then 'Feb'
when 3 then 'Mar'
when 4 then 'Apr'
when 5 then 'May'
when 6 then 'Jun'
when 7 then 'Jul'
when 8 then 'Aug'
when 9 then 'Sep'
when 10 then 'Oct'
when 11 then 'Nov'
when 12 then 'Dev'
else '***'
end ,
member_cnt = sum( case when m.member_number is not null then 1 else 0 end )
from ( select yyyy = yyyy.value ,
mm = mm.value ,
dtFrom = dateadd( month , mm.value - 1 , dateadd( year , yyyy.value - 1900 , convert(date,'') ) ) ,
dtThru = dateadd( day , - 1 , dateadd( month , mm.value , dateadd( year , yyyy.value - 1900 , convert(date,'') ) ) )
from dbo.IntegerRange(2000,2013) yyyy
full join dbo.IntegerRange(1,12) mm on 1=1
) period
left join dbo.members m on period.dtFrom <= m.end_date
and period.dtThru >= m.start_date
group by period.yyyy ,
period.mm
order by period.yyyy ,
period.mm
The first table expression in the from clause creates a virtual table of the periods (months, in this case, but the technique doesn't limit itself to months or even weeks) covering the reporting period:
from ( select yyyy = yyyy.value ,
mm = mm.value ,
dtFrom = dateadd( month , mm.value - 1 , dateadd( year , yyyy.value - 1900 , convert(date,'') ) ) ,
dtThru = dateadd( day , - 1 , dateadd( month , mm.value , dateadd( year , yyyy.value - 1900 , convert(date,'') ) ) )
from dbo.IntegerRange(2000,2013) yyyy
full join dbo.IntegerRange(1,12) mm on 1=1
) period
That is then joined, via a left outer join, ensuring that all periods are reported, not just those periods with active members, to the members table to collect, for each reporting period in the virtual period table above, the set of members who were active during the period:
left join dbo.members m on period.dtFrom <= m.end_date
and period.dtThru >= m.start_date
We then group by the year and month of each period and then order the results by year/month number:
group by period.yyyy ,
period.mm
order by period.yyyy ,
period.mm
In creating the results set to be returned, we return the year of the period, the month number (converted to a friendly name), and the count of active members. Note that we have to use the sum() aggregate function here rather than count() as empty periods will have a single row returned (with null in all columns). Count(), unlike all other aggregate functions, includes null values in the aggregation. Sum() is applied to a case expression acting as a discriminant function returning 1 or 0 identifying whether the row indicates useful or missing data:
select [year] = period.yyyy ,
[month] = case period.mm ,
when 1 then 'Jan'
when 2 then 'Feb'
when 3 then 'Mar'
when 4 then 'Apr'
when 5 then 'May'
when 6 then 'Jun'
when 7 then 'Jul'
when 8 then 'Aug'
when 9 then 'Sep'
when 10 then 'Oct'
when 11 then 'Nov'
when 12 then 'Dev'
else '***'
end ,
member_cnt = sum( case when m.member_number is not null then 1 else 0 end )
Easy!
DECLARE #minMonth DATE
SELECT #minMonth = MIN(StartDate) FROM Table1
DECLARE #maxMonth DATE
SELECT #maxMonth = MAX(EndDate) FROM Table1
;WITH CTE_Months AS
(
SELECT #minMonth AS Mnth
UNION ALL
SELECT DATEADD(MM,1,Mnth) FROM CTE_Months
WHERE Mnth<#MaxMonth
)
SELECT Mnth AS Month, COUNT(*) as Members
FROM CTE_Months m
LEFT JOIN Table1 t on m.Mnth BETWEEN t.StartDate AND t.EndDate
GROUP BY Mnth
SQLFiddle Demo
CTE will find all months from min StartDate to max EndDate, if you need different min and max, just change how you get #MinMonth and #MaxMonth
If you don't want to show zeros for months that possibly don't have any members, replace LEFT JOIN with INNER at the end.
select t.Month, count(*) Members
from (
select case
when startdate <= to_date('01-Jan-2002', 'DD-MON-YYYY') AND enddate >= to_date('31-Jan-2002', 'DD-MON-YYYY') then ' Jan'
when startdate <= to_date('01-Feb-2002', 'DD-MON-YYYY') AND enddate >= to_date('28-Feb-2002', 'DD-MON-YYYY') then ' Feb'
when startdate <= to_date('01-Mar-2002', 'DD-MON-YYYY') AND enddate >= to_date('31-Mar-2002', 'DD-MON-YYYY') then ' Mar'
when startdate <= to_date('01-Apr-2002', 'DD-MON-YYYY') AND enddate >= to_date('30-Apr-2002', 'DD-MON-YYYY') then ' Apr'
when startdate <= to_date('01-May-2002', 'DD-MON-YYYY') AND enddate >= to_date('31-May-2002', 'DD-MON-YYYY') then ' May'
when startdate <= to_date('01-Jun-2002', 'DD-MON-YYYY') AND enddate >= to_date('30-Jun-2002', 'DD-MON-YYYY') then ' Jun'
when startdate <= to_date('01-Jul-2002', 'DD-MON-YYYY') AND enddate >= to_date('31-Jul-2002', 'DD-MON-YYYY') then ' Jul'
when startdate <= to_date('01-Aug-2002', 'DD-MON-YYYY') AND enddate >= to_date('31-Aug-2002', 'DD-MON-YYYY') then ' Aug'
when startdate <= to_date('01-Sep-2002', 'DD-MON-YYYY') AND enddate >= to_date('30-Sep-2002', 'DD-MON-YYYY') then ' Sep'
when startdate <= to_date('01-Oct-2002', 'DD-MON-YYYY') AND enddate >= to_date('31-Oct-2002', 'DD-MON-YYYY') then ' Oct'
when startdate <= to_date('01-Nov-2002', 'DD-MON-YYYY') AND enddate >= to_date('30-Nov-2002', 'DD-MON-YYYY') then ' Nov'
when startdate <= to_date('01-Dec-2002', 'DD-MON-YYYY') AND enddate >= to_date('31-Dec-2002', 'DD-MON-YYYY') then ' Dec'
end as Month
from member) t
group by t.Month
Where member is the table name.
select
extract(year from datetimestamp ) Yr,extract(month from datetimestamp) Mn,
c.weekday_of_month wk, a.aircraft_type,count( a.aircraft_type) from fcm_bv.Flights b
join fcm_bv.Fleet a on b.aircraftid=a.tail
join SYS_CALENDAR.CALENDAR c
on cast(b.datetimestamp AS DATE FORMAT 'YYYY-MM-DD') = cast(c.calendar_date AS DATE FORMAT 'YYYY-MM-DD')
where cast(datetimestamp as date) >= '2011-09-01'
and cast(datetimestamp as date) <= '2011-09-30' order by wk
group by Yr,Mn,wk,a.fleet,a.aircraft_type
While Running above Query I am getting out put like this
Yr Mn wk AIRCRAFT_TYPE Count(AIRCRAFT_TYPE)
2011 9 1 B737-700 1744
2011 9 1 B737-800 131
2011 9 1 B737-800W 2711
2011 9 1 B737-8BK 180
2011 9 1 B737-700W 329
But I need output in below format
Yr Mn wk B737-700 B737-800 B737-800W B737-8BK B737-700W
2011 9 1 1744 131 2711 180 329
Could any one help me
In the past when I have needed to do this the pivot was against a discrete, managable volume of categories and the following SQL has served me well:
SELECT EXTRACT(YEAR FROM b.datetimestamp) AS Yr
, EXTRACT(MONTH FROM b.datetimestamp) AS Mn
, C.weekday_of_month
, COUNT(CASE WHEN a.aircraft_type = 'B737-700' THEN a.aircraft_type ELSE NULL END) AS B737-700
, COUNT(CASE WHEN a.aircraft_type = 'B737-800' THEN a.aircraft_type ELSE NULL END) AS B737-800
, /* Other Known Aircrafts */
, COUNT(CASE WHEN a.aircrat_type NOT IN ('<list of known aircraft types>') THEN a.aircraft_type ELSE NULL END) AS Uncategorized_Aircraft
FROM fcm_bv.Flights b
join fcm_bv.Fleet a on b.aircraftid=a.tail
join SYS_CALENDAR.CALENDAR c
on cast(b.datetimestamp AS DATE FORMAT 'YYYY-MM-DD') = cast(c.calendar_date AS DATE FORMAT 'YYYY-MM-DD')
WHERE cast(datetimestamp as date) >= '2011-09-01'
AND cast(datetimestamp as date) <= '2011-09-30' order by wk
GROUP BY Yr,Mn,wk,a.fleet
If you have to pivot against a constantly changing category it may be best to leave the pivoting to MS Excel or BI tool of choice.
Hope this helps.