I have re-written this query about 20 times today and I keep getting close but no dice... I'm sure this is easy-peasy for y'all, but my SQL (Oracle) is pretty rusty.
Here's what I need:
PersonID Count1 Count2 Count3 Count4
1 0 0 2 1
2 1 1 1 0
3 1 1 1 2
Data is coming from several sources. I have a table People, and a table Values. People can have any number of values in that table.
PersonID Item Value
1 Check1 3
1 Check2 3
1 Check3 4
2 Check4 2
2 Check5 3
2 Check6 1
.. etc
So the query would, for each PersonID, count how many times the particular Value appears. The values are always 1, 2, 3, or 4. I tried to do 4 subqueries, but it wouldn't read the PersonID from the main query and just returned the count of all instances of value=1.
I was then thinking do a Group_By ... I don't know. Any help is appreciated!
ETA: I've deleted & re-written the query many times in many ways and unfortunately did not save any intermediate attempts. I didn't include it originally because I was in the middle of rearranging it again, and it's not runnable as-is. But here it is as it stands now:
/*sources are the tested requirements
values are the scores people received on the tested sources
people are those who were tested on the requirements */
WITH sub_query4 (
SELECT values.personid,
count (values.ID) as count4 --how many 4s
FROM values
INNER JOIN sources ON values.valueid = sources.sourceid
INNER JOIN people ON people.personid = values.personid
WHERE values.yearid = 2017
AND values.quarter = 'Q1'
AND instr (sources.identifier, 'TESTBANK.01', 1 ,1) > 0
AND values.value = '4'
GROUP_BY people.personid
)
SELECT p.first_name,
p.last_name,
p.position,
p.email,
p.locationid,
sub_query4.count4 as count4 --eventually this would repeat for 1, 2, & 3
FROM people p
WHERE p.locationid=406
AND p.position in (9,10);
values is a bad name for a table because it is a SQL keyword.
In any case, conditional aggregation should work:
select personid,
sum(case when value = 1 then 1 else 0 end) as cnt_1,
sum(case when value = 2 then 1 else 0 end) as cnt_2,
sum(case when value = 3 then 1 else 0 end) as cnt_3,
sum(case when value = 4 then 1 else 0 end) as cnt_4
from values
group by personid;
I prefer to use PIVOT for this. Here is Example SQL Fiddle
SELECT "PersonID", val1,val2,val3,val4 FROM
(
SELECT "PersonID", "Value" from VALS
)
PIVOT
(
count("Value")
FOR "Value" IN (1 as val1, 2 as val2, 3 as val3, 4 as val4)
);
Related
Consider the following tables
ID_TYPE
ID
TYPE
1
3
2
3
3
1
4
2
5
2
ID_HISTORY
DEBIT_ID
DEBIT_LOCATION
AMOUNT
CREDIT_ID
CREDIT_LOCATION
MONTH
3
LOC1
100
1
LOC5
MAY
4
LOC2
200
3
LOC6
MAY
2
LOC3
300
5
LOC7
MAY
1
LOC4
400
3
LOC8
JUNE
3
LOC9
500
2
LOC10
JUNE
Now suppose I want to fetch all rows from ID_HISTORY in the MONTH of MAY, and result should contain only these columns:
Id, Location, Amount
Cases:
Result should contain rows only where either DEBIT_ID or CREDIT_ID is of TYPE=3 in ID_TYPE table
If the DEBIT_ID is of TYPE = 3 in the ID_TYPE table, then pick DEBIT_ID as "Id", else pick CREDIT_ID as "Id"
Similarly, If the DEBIT_ID is of TYPE 3 in the ID_TYPE table, then pick DEBIT_LOCATION as "Location", else pick CREDIT_LOCATION as "Location"
For example, above tables should result in the following:
Id
Location
Amount
1
LOC5
100
2
LOC3
300
I know that something like the following should work:
SELECT
(CASE
WHEN (Tab.DEBIT_ID IN (
SELECT ID
FROM ID_TYPE Typ
WHERE Typ.TYPE = 3)
) THEN Tab.DEBIT_ID
ELSE Tab.CREDIT_ID END
) "Id",
(CASE
WHEN (Tab.DEBIT_ID IN (
SELECT ID
FROM ID_TYPE Typ
WHERE Typ.TYPE = 3)
) THEN Tab.DEBIT_LOCATION
ELSE Tab.CREDIT_LOCATION END
) "Location",
Tab.AMOUNT "Amount"
FROM (
SELECT *
FROM ID_HISTORY Tab
WHERE Tab.MONTH = 'MAY'
--this block will be very complicated and contain complex multi-level queries to fetch data
)
But as you can see this will be inefficient as I have to basically duplicate the full case logic for each conditional columns. Also, this is no way "clean" in case there are a lot of similar columns.
Also, if the case logic is complex, it will be inefficient even further. It would be better if i could select multiple columns in THEN / ELSE cases. I tried doing that, but it just gives me "ORA-00913: too many values" error.
What would be the optimized version?
You could use a join to remove the sub-queries:
SELECT CASE
WHEN typ.id IS NOT NULL
THEN h.debit_id
ELSE h.credit_id
END AS id,
CASE
WHEN typ.id IS NOT NULL
THEN h.debit_location
ELSE h.credit_location
END AS location,
h.AMOUNT
FROM (
SELECT *
FROM ID_HISTORY
WHERE MONTH = 'MAY'
-- this block will be very complicated and contain complex multi-level queries to fetch data
) h
LEFT OUTER JOIN (
SELECT id
FROM id_type
WHERE type = 3
) typ
ON (h.debit_id = typ.id)
I have a table grades[Student, Grade]
My task is to сount the amount of students, have 2 A's (2) and at least one D (2).
The most painful part is that i shouldn't use sql subqueries
That abomination i come up with would only list such nulls and students which fit by condition
Any ideas how to count them ? Or rework query?
SELECT CASE WHEN
(
sum(CASE WHEN g1.grade = 5
THEN 1
ELSE 0
END) > 2
)
AND
(
sum(CASE WHEN g1.grade = 2
THEN 1
ELSE 0
END) > 0
)
THEN g1.student_name
ELSE NULL
END as student_name
FROM grades g1
group by g1.student_name
You don't need any sub queries you can just count and filter using having and conditional sum
select Student
from Grades
group by Student
having sum(case when Grade='A' then 1 end)=2
and sum(case when Grade='D' then 1 end)>0
Edit
To get just the count of qualifying rows you can do
select distinct count(*) over()
from Grades
group by Student
having sum(case when Grade='A' then 1 end)=2
and sum(case when Grade='D' then 1 end)>0
Oh how I miss homework....
Depending on your database, you can use PIVOT:
SELECT *
FROM grades PIVOT (COUNT(*) AS CNT FOR grade IN (5 AS A, 2 AS D)) g
WHERE g.a_cnt >= 2 AND g.d_cnt > 0;
Alternate:
SELECT g.student
FROM grades g
GROUP BY g.student
HAVING COUNT(DECODE(g.grade, 5, 1, NULL)) >= 2 AND COUNT(DECODE(g.grade, 2, 1, NULL)) > 0;
TSQL - SQL Sever
I'm building a report to very specific requirements. I'm trying to combine multiples of a key ID into single rows, but there's different values in some of the columns, so GROUP BY won't work.
SELECT count(tt.Person_ID) as CandCount, tt.Person_ID,
CASE e.EthnicSuperCategoryID WHEN CandCount > 1 THEN 10 ELSE e.EthnicSuperCategoryID END as EthnicSuperCategoryID,
CASE e.Ethnicity_Id WHEN 1 THEN 1 ELSE 0 END as Black ,
CASE e.Ethnicity_Id WHEN 2 THEN 1 ELSE 0 END as White ,
CASE e.Ethnicity_Id WHEN 3 THEN 1 ELSE 0 END as Asian,
etc
FROM T_1 TT
JOINS
WHERE
GROUP
Msg 102, Level 15, State 1, Line 4
Incorrect syntax near '>'.
Here's the results (without the first CASE). Note person 3 stated multiple ethnicities.
SELECT count(tt.Person_ID) as CandCount, tt.Person_ID,
CASE e.Ethnicity_Id WHEN 1 THEN 1 ELSE 0 END as Black ,
CASE e.Ethnicity_Id WHEN 2 THEN 1 ELSE 0 END as White ,
CASE e.Ethnicity_Id WHEN 3 THEN 1 ELSE 0 END as Asian,
etc
FROM T_1 TT
JOINS
WHERE
GROUP
That’s expected, but the goal would be to assign multiple ethnicities to Ethnicity_Id of 10 (multiple). I also want them grouped on a single line.
So the end result would look like this:
So my issue is two fold. If the candidate has more than 2 ethnicities, assign the records to Ethnicity_Id of 10. I also need duplicated person IDs grouped into a single row, while displaying all of the results of the columns.
This should bring your desired result:
SELECT Person_ID
, ISNULL(ID_Dummy,Ethnicity_ID) Ethnicity_ID
, MAX(Black) Black
, MAX(White) White
, MAX(Asian) Asian
FROM #T T
OUTER APPLY(SELECT MAX(10) FROM #T T2
WHERE T2.Person_ID = T.Person_ID
AND T2.Ethnicity_ID <> T.Ethnicity_ID
)EthnicityOverride(ID_Dummy)
GROUP BY Person_ID, ISNULL(ID_Dummy,Ethnicity_ID)
You want conditional aggregation. Your query is incomplete, but the idea is:
select
person_id,
sum(case ethnicity_id = 1 then 1 else 0 end) as black,
sum(case ethnicity_id = 2 then 1 else 0 end) as white,
sum(case ethnicity_id = 3 then 1 else 0 end) as asian
from ...
where ...
group by person_id
You might want max() instead of sum(). Also I did not get the logic for column the second column in the desired results - maybe that's just count(*).
This would be my approach
SELECT
person_id,
CASE WHEN flag = 1 THEN Ethnicity_Id ELSE 10 END AS Ethnicity_Id,
[1] as black,
[2] as white,
[3] as asian
FROM
(
SELECT
person_id,
Ethnicity_Id as columns,
1 as n,
MAX(Ethnicity_Id) over(PARTITION BY person_id) as Ethnicity_Id,
COUNT(Ethnicity_Id) over(PARTITION BY person_id) as flag
FROM
#example
) AS SourceTable
PIVOT
(
MAX(n) FOR columns IN ([1], [2], [3])
) AS PivotTable;
Pivot the Ethnicity_Id column into multiples columns, Using constant
1 to make it complain with your expected result.
Using Max(Ethnicity_Id) with Partition By to get the original
Ethnicity_Id
Using Count(Ethnicity_Id) to flag if a need to raplace Ethnicity_Id
with 10 bc there is more that 1 row for that person_id
If you need to add more Ethnicitys add the ids in ... IN ([1], [2], [3]) ... and in the select
Quick one,
I have a table, with the following structure
id lid taken
1 1 0
1 1 0
1 1 1
1 1 1
1 2 1
Pretty simply so far right?
I need to query the taken/available from the lid of 1, which should return
taken available
2 2
I know I can simply do two counts and join them, but is there a more proficient way of doing this rather than two separate queries?
I was looking at the following type of format, but I can not for the life of me get it executed in SQL...
SELECT
COUNT(case taken=1) AS taken,
COUNT(case taken=0) AS available FROM table
WHERE
lid=1
Thank you SO much.
You can do this:
SELECT taken, COUNT(*) AS count
FROM table
WHERE lid = 1
GROUP BY taken
This will return two rows:
taken count
0 2
1 2
Each count corresponds to how many times that particular taken value was seen.
Your query is correct just needs juggling a bit:
SELECT
SUM(case taken WHEN 1 THEN 1 ELSE 0 END) AS taken,
SUM(case taken WHEN 1 THEN 0 ELSE 1 END) AS available FROM table
WHERE
lid=1
Alternatively you could do:
SELECT
SUM(taken) AS taken,
COUNT(id) - SUM(taken) AS available
FROM table
WHERE
lid=1
SELECT
SUM(case WHEN taken=1 THEN 1 ELSE 0 END) AS taken,
SUM(case WHEN taken=0 THEN 1 ELSE 0 END) AS available
FROM table
WHERE lid=1
Weird application of CTE's:
WITH lid AS (
SELECT DISTINCT lid FROM taken
)
, tak AS (
SELECT lid,taken , COUNT(*) AS cnt
FROM taken t0
GROUP BY lid,taken
)
SELECT l.lid
, COALESCE(a0.cnt, 0) AS available
, COALESCE(a1.cnt, 0) AS taken
FROM lid l
LEFT JOIN tak a0 ON a0.lid=l.lid AND a0.taken = 0
LEFT JOIN tak a1 ON a1.lid=l.lid AND a1.taken = 1
WHERE l.lid=1
;
My table structure is this
id last_mod_dt nr is_u is_rog is_ror is_unv
1 x uuid1 1 1 1 0
2 y uuid1 1 0 1 1
3 z uuid2 1 1 1 1
I want the count of rows with:
is_ror=1 or is_rog =1
is_u=1
is_unv=1
All in a single query. Is it possible?
The problem I am facing is that there can be same values for nr as is the case in the table above.
Case statments provide mondo flexibility...
SELECT
sum(case
when is_ror = 1 or is_rog = 1 then 1
else 0
end) FirstCount
,sum(case
when is_u = 1 then 1
else 0
end) SecondCount
,sum(case
when is_unv = 1 then 1
else 0
end) ThirdCount
from MyTable
you can use union to get multiple results e.g.
select count(*) from table with is_ror=1 or is_rog =1
union
select count(*) from table with is_u=1
union
select count(*) from table with is_unv=1
Then the result set will contain three rows each with one of the counts.
Sounds pretty simple if "all in a single query" does not disqualify subselects;
SELECT
(SELECT COUNT(DISTINCT nr) FROM table1 WHERE is_ror=1 OR is_rog=1) cnt_ror_reg,
(SELECT COUNT(DISTINCT nr) FROM table1 WHERE is_u=1) cnt_u,
(SELECT COUNT(DISTINCT nr) FROM table1 WHERE is_unv=1) cnt_unv;
how about something like
SELECT
SUM(IF(is_u > 0 AND is_rog > 0, 1, 0)) AS count_something,
...
from table
group by nr
I think it will do the trick
I am of course not sure what you want exactly, but I believe you can use the logic to produce your desired result.