I want to simply add a column to the results given below that comes from "select count(*) ... group by possession". So it should still retain the same number of rows, and have this added column. I was told to look into a lateral join but I don't understand how to do it especially in the context of my query which has the CTE
QUERY
select
*
from (
with possession_change as (
select
(lag(possession,1) over (order by id)) as last_possession,
possession,
clock
from plays
where
game_id in (583615)
and league = 3
and period in (0,1)
)
select * from possession_change
) stuff
;
RESULTS
last_possession | possession | clock
-----------------+------------+-------
| 0 | 3600
0 | 0 | 3600
0 | 0 | 3600
0 | 0 | 3600
0 | 1 | 3561
1 | 1 | 3561
1 | 1 | 3561
1 | 1 | 3449
1 | 1 | 3449
1 | 0 | 3396
0 | 0 | 3396
0 | 0 | 3396
DESIRED RESULTS
last_possession | possession | clock | possession_count
-----------------+------------+-------
| 0 | 3600 | 7
0 | 0 | 3600 | 7
0 | 0 | 3600 | 7
0 | 0 | 3600 | 7
0 | 1 | 3561 | 5
1 | 1 | 3561 | 5
1 | 1 | 3561 | 5
1 | 1 | 3449 | 5
1 | 1 | 3449 | 5
1 | 0 | 3396 | 7
0 | 0 | 3396 | 7
0 | 0 | 3396 | 7
You can use count over:
select
lag(possession,1) over (order by id) as last_possession,
possession,
clock,
count(*) over (partition by possession) cnt
from plays
where
game_id in (583615)
and league = 3
and period in (0,1)
Related
I have a table of Account Transactions that includes ID, Amount, Date. Basically, I want to create a resulting table that looks at the table and returns what the SUM was for the Account over three different Ending Date Ranges. Then I want to Flag (Combined_Flag) each Account ID, 1 if any of the SUMs for that ID are non-zero, and a 0 if all of the SUMs are 0.
Date Range 1) Min Date to End of Last Month (-1 Month)
Date Range 2) Min Date to End of 2 Months ago (-2 Months)
Date Range 3) Min Date to End of Last Month, Last Year (-13 Months)
The Resulting table should be: ID, SUM_R1, SUM_R2, SUM_R3, Flag_R1, Flag_R2, Flag_R3, Combined_Flag
Example Data
| ID | Amount | Date |
| -------- | -------------- |-------------- |
| 1 | 20 | 09/01/19 |
| 2 | 40 | 09/01/19 |
| 3 | 0 | 09/01/19 |
| 4 | 0 | 09/01/19 |
| 1 | 10 | 10/01/19 |
| 2 | 0 | 10/01/19 |
| 3 | 0 | 10/01/19 |
| 4 | 0 | 10/01/19 |
| 1 | 15 | 11/01/19 |
| 2 | 40 | 11/01/19 |
| 3 | 0 | 11/01/19 |
| 4 | 0 | 11/01/19 |
| 1 | 20 | 09/01/20 |
| 2 | 40 | 09/01/20 |
| 3 | 0 | 09/01/20 |
| 4 | 50 | 09/01/20 |
| 1 | 10 | 10/01/20 |
| 2 | 0 | 10/01/20 |
| 3 | 0 | 10/01/20 |
| 4 | 65 | 10/01/20 |
| 1 | 15 | 11/01/20 |
| 2 | 40 | 11/01/20 |
| 3 | 0 | 11/01/20 |
| 4 | 0 | 11/01/20 |
Expected Result Table (Using Date of 12/21/2020)
| ID | SUM_R1 | SUM_R2 | SUM_R3 | Flag_R1 | Flag_R2 | Flag_R3 | Combined_Flag |
| -------- | -------- | -------- | -------- | --------- | --------- | --------- | --------------- |
| 1 | 90 | 75 | 45 | 1 | 1 | 1 | 1 |
| 2 | 160 | 120 | 80 | 1 | 1 | 1 | 1 |
| 3 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 4 | 115 | 115 | 0 | 1 | 1 | 0 | 1 |
The difficulty I'm having is in joining the table basically to itself 2 times. I'm getting results all over the place and not really sure exactly what's going on.
Is this what you want?
select id,
sum(case when date < datefromparts(year(v.dte), month(v.dte), 1)
then amount else 0
end) as sum_r1,
sum(case when date < dateadd(month, -1, datefromparts(year(v.dte), month(v.dte), 1))
then amount else 0
end) as sum_r2,
sum(case when date < dateadd(month, -13, datefromparts(year(v.dte), month(v.dte), 1))
then amount else 0
end) as sum_r3,
max(case when amount > 0 and date < datefromparts(year(v.dte), month(v.dte), 1)
then 1 else 0
end) as flag_r1,
max(case when amount > 0 and date < dateadd(month, -1, datefromparts(year(v.dte), month(v.dte), 1))
then 1 else 0
end) as flag_r2,
max(case when amount > 0 and date < dateadd(month, -13, datefromparts(year(v.dte), month(v.dte), 1))
then 1 else 0
end) as flag_r3
from t cross join
(values (convert(date, '2020-12-21'))
) v(dte)
group by id;
The flag columns assume that the amounts are never negative (which is consistent with the data in your question.
EDIT:
The shorthand in the comment for creating the flag looks like:
abs(sign(sum(case when amount > 0 and date < datefromparts(year(v.dte), month(v.dte), 1)
then amount else 0
end))) as flag_r1,
Here is a db<>fiddle.
Suppose I have a table like this:
id | part | value
----+-------+-------
1 | 0 | 8
2 | 0 | 3
3 | 0 | 4
4 | 1 | 6
5 | 0 | 13
6 | 0 | 4
7 | 1 | 2
8 | 0 | 11
9 | 0 | 15
10 | 0 | 3
11 | 0 | 2
I would like to enumerate groups that have part atribute 0.
Ultimately I want to get this:
id | part | value | number
----+-------+-----------------
1 | 0 | 8 | 1
2 | 0 | 3 | 2
3 | 0 | 4 | 3
4 | 1 | 6 | 0
5 | 0 | 13 | 1
6 | 0 | 4 | 2
7 | 1 | 2 | 0
8 | 0 | 11 | 1
9 | 0 | 15 | 2
10 | 0 | 3 | 3
11 | 0 | 2 | 4
Is it possible to solve this with Postgres window functions or is there another way?
Yes, that is simple:
SELECT id, part, value,
row_number() OVER (PARTITION BY grp ORDER BY id) - 1 AS number
FROM (SELECT id, part, value,
sum(part) OVER (ORDER BY id) AS grp
FROM mytable
) AS q;
id | part | value | number
----+------+-------+--------
1 | 0 | 8 | 0
2 | 0 | 3 | 1
3 | 0 | 4 | 2
4 | 1 | 6 | 0
5 | 0 | 13 | 1
6 | 0 | 4 | 2
7 | 1 | 2 | 0
8 | 0 | 11 | 1
9 | 0 | 15 | 2
10 | 0 | 3 | 3
11 | 0 | 2 | 4
(11 rows)
My table returns results as following (skips row if HourOfDay does not have data for particular ID)
ID HourOfDay Counts
--------------------------
1 5 5
1 13 10
1 23 3
..........................HourOfDay up till 23
2 9 1
and so on.
What I am trying to achieve is to force showing rows displaying 0 for HoursOfDay, which don't have data, like following:
ID HourOfDay Counts
--------------------------
1 0 0
1 1 0
1 2 0
1......................
1 5 5
1 6 0
1......................
1 23 3
2 0 0
2 1 0
etc.
I have researched around about it. It looks like I can achieve this result if I create an extra table and outer join it. So I have created table variable in SP (as a temp workaround)
DECLARE #Hours TABLE
(
[Hour] INT NULL
);
INSERT INTO #Hours VALUES (0),(1),(2),(3),(4),(5),(6),(7),(8),(9),(10),(11),(12)
,(13),(14),(15),(16),(17),(18),(19),(20),(21),(22),(23);
However, no matter how I join it, it does not achieve desired result.
How do I proceed? Do I add extra columns to join on? Completely different approach? Any hint in the right direction is appreciated!
Using a derived table for the distinct Ids cross joined to #Hours, left joined to your table:
select
i.Id
, h.Hour
, coalesce(t.Counts,0) as Counts
from (select distinct Id from t) as i
cross join #Hours as h
left join t
on i.Id = t.Id
and h.Hour = t.HourOfDay
rextester demo: http://rextester.com/XFZYX88502
returns:
+----+------+--------+
| Id | Hour | Counts |
+----+------+--------+
| 1 | 0 | 0 |
| 1 | 1 | 0 |
| 1 | 2 | 0 |
| 1 | 3 | 0 |
| 1 | 4 | 0 |
| 1 | 5 | 5 |
| 1 | 6 | 0 |
| 1 | 7 | 0 |
| 1 | 8 | 0 |
| 1 | 9 | 0 |
| 1 | 10 | 0 |
| 1 | 11 | 0 |
| 1 | 12 | 0 |
| 1 | 13 | 10 |
| 1 | 14 | 0 |
| 1 | 15 | 0 |
| 1 | 16 | 0 |
| 1 | 17 | 0 |
| 1 | 18 | 0 |
| 1 | 19 | 0 |
| 1 | 20 | 0 |
| 1 | 21 | 0 |
| 1 | 22 | 0 |
| 1 | 23 | 3 |
| 2 | 0 | 0 |
| 2 | 1 | 0 |
| 2 | 2 | 0 |
| 2 | 3 | 0 |
| 2 | 4 | 0 |
| 2 | 5 | 0 |
| 2 | 6 | 0 |
| 2 | 7 | 0 |
| 2 | 8 | 0 |
| 2 | 9 | 1 |
| 2 | 10 | 0 |
| 2 | 11 | 0 |
| 2 | 12 | 0 |
| 2 | 13 | 0 |
| 2 | 14 | 0 |
| 2 | 15 | 0 |
| 2 | 16 | 0 |
| 2 | 17 | 0 |
| 2 | 18 | 0 |
| 2 | 19 | 0 |
| 2 | 20 | 0 |
| 2 | 21 | 0 |
| 2 | 22 | 0 |
| 2 | 23 | 0 |
+----+------+--------+
I have a table which has 7 different classes with an area value.
pid | class| area |
----+------+------+
2 | 1 | 10 |
2 | 2 | 10 |
2 | 6 | 20 |
4 | 1 | 30 |
4 | 2 | 40 |
4 | 3 | 50 |
4 | 4 | 60 |
4 | 5 | 70 |
9 | 6 | 80 |
11 | 1 | 90 |
11 | 4 | 10 |
11 | 7 | 20 |
However I want to present this data in a format that has each distinct pid as a column heading and then have each row correspond to a class area (i.e. first row is the area of class 1 for each pid).
2 | 4 | 9 | 11 |
---+-----+-----+----+
10 | 30 | 0 | 90 |
10 | 40 | 0 | 0 |
0 | 50 | 0 | 0 |
0 | 60 | 0 | 10 |
0 | 70 | 0 | 0 |
20 | 0 | 60 | 0 |
0 | 0 | 0 | 20 |
Is it possible to create an output like this in PostgreSQL?
Try this:
SELECT
SUM(CASE WHEN pid = 2 THEN area ELSE 0 END) As "2",
SUM(CASE WHEN pid = 4 THEN area ELSE 0 END) As "4",
SUM(CASE WHEN pid = 9 THEN area ELSE 0 END) As "9",
SUM(CASE WHEN pid = 11 THEN area ELSE 0 END) As "11"
FROM t
GROUP BY class
ORDER BY class
I am trying to subtract two columns and then count how many have the same difference and put those sums into columns. The sums are of how many have a difference of -3 or more, -2, -1, 0, 1, 2, 3 or more grouped by date.
Query must be executed against a DB2 database.
Data...
------------------------------
| Date | Num 1 | Num 2 |
------------------------------
| 2014-02-11 | 19872 | 19873 |
| 2014-02-11 | 19873 | 19873 |
| 2014-02-12 | 19875 | 19873 |
| 2014-02-13 | 19870 | 19873 |
| 2014-02-13 | 19872 | 19873 |
| 2014-02-14 | 19877 | 19869 |
| 2014-02-14 | 19873 | 19873 |
Desired Output...
-----------------------------------------------------------------------
| Date | <= -3 | -2 | -1 | 0 | +1 | +2 | >= +3 |
-----------------------------------------------------------------------
| 2014-02-11 | 0 | 0 | 0 | 1 | 1 | 0 | 0 |
| 2014-02-12 | 0 | 1 | 0 | 0 | 0 | 0 | 0 |
| 2014-02-13 | 0 | 0 | 0 | 0 | 1 | 0 | 1 |
| 2014-02-14 | 1 | 0 | 0 | 1 | 9 | 0 | 0 |
Try this:
select Date,
sum(case when diff <= -3 then 1 else 0 ) AS [<=-3],
sum(case when diff = -2 then 1 else 0 ) AS [-2],
sum(case when diff = -1 then 1 else 0 ) AS [-1],
sum(case when diff = 0 then 1 else 0 ) AS [0],
sum(case when diff = 1 then 1 else 0 ) AS [+1],
sum(case when diff = 2 then 1 else 0 ) AS [+2],
sum(case when diff >= 3 then 1 else 0 ) AS [>=+3]
from
(select Date, Num1, Num2, (Num1-Num2) diff from TableA)TableB
group by Date