How to do mathematical comparision with SQL CASE statement - sql

I have a student table around 100k records and I have two types of data in it: student name and level type with selection values primary, secondary, intermediate & university
I want to filter out the student from this table, whose have count > 0, in all level primary, secondary, intermediate & university
I was able to find the sum for each student in each level using the following query
SELECT
student_id,
SUM(CASE WHEN lev_type = 'primary' THEN 1 ELSE 0 END) AS primary,
SUM(CASE WHEN lev_type = 'secondary' THEN 1 ELSE 0 END) AS secondary,
SUM(CASE WHEN lev_type = 'intermediate' THEN 1 ELSE 0 END) AS intermediate,
SUM(CASE WHEN level_type = 'university' THEN 1 ELSE 0 END) AS university
FROM
student_details
GROUP BY
student_id
and I am getting a result like (note that my result is 92242 row(s))
attendee_id primary secondary intermediate uni
student1 0 1 1 2
student2 0 1 1 0
student3 88 209 92 32
student4 0 1 1 0
student5 0 1 1 0
How to filter out student3 from this result?

You can simply add a where statement as follows:
SELECT student_id,
SUM(case when lev_type = 'primary' then 1 else 0 end) as primary,
SUM(case when lev_type = 'secondary' then 1 else 0 end) as secondary ,
SUM(case when lev_type = 'intermediate' then 1 else 0 end) as intermediate ,
SUM(case when lev_type = 'university' then 1 else 0 end) as university
FROM student_details
GROUP BY student_id
WHERE primary = 0 OR secondary = 0 OR intermediate = 0 OR university = 0

HAVING might get you what you want. For example:
SELECT student_id,
sum(case when lev_type = 'primary' then 1 else 0 end) as primary,
sum(case when lev_type = 'secondary' then 1 else 0 end) as secondary ,
sum(case when lev_type = 'intermediate' then 1 else 0 end) as intermediate ,
sum(case when level_type = 'university' then 1 else 0 end) as university
from student_details
group by student_id
-- Put the criteria here by which you want to filter
having sum(case when lev_type = 'primary' then 1 else 0 end) = 0
and sum(case when lev_type = 'secondary' then 1 else 0 end) = 0
and sum(case when lev_type = 'intermediate' then 1 else 0 end) = 0
and sum(case when level_type = 'university' then 1 else 0 end) = 0

Related

Count the number of column having non zero values

I have the following script,
SELECT COUNT(*) AS Total,
SUM(CASE WHEN TypeId ='4' THEN 1 ELSE 0 END) AS 'TotalCount1',
SUM(CASE WHEN TypeId ='6' THEN 1 ELSE 0 END) AS 'TotalCount2',
SUM(CASE WHEN TypeId ='1' THEN 1 ELSE 0 END) AS 'TotalCount3',
SUM(CASE WHEN TypeId ='10' THEN 1 ELSE 0 END) AS 'TotalCount4',
SUM(CASE WHEN TypeId ='5' THEN 1 ELSE 0 END) AS 'TotalCount5',
SUM(CASE WHEN TypeId ='8' THEN 1 ELSE 0 END) AS 'TotalCount6'
FROM [Party]
Please refer the screenshot as the output of the above script.
What I want:
I want a column after the Total as the total number of the column having nonzero values.
Like in the picture the values should be 2 as TotalCount1 and Totalcount3 have non zero values.
SELECT COUNT(*) AS Total,
...
...
CASE ( WHEN SUM(CASE WHEN TypeId ='4' THEN 1 ELSE 0 END) > 0 THEN 1 ELSE 0 END) +
CASE ( WHEN SUM(CASE WHEN TypeId ='6' THEN 1 ELSE 0 END) > 0 THEN 1 ELSE 0 END) +
CASE ( WHEN SUM(CASE WHEN TypeId ='1' THEN 1 ELSE 0 END) > 0 THEN 1 ELSE 0 END) +
CASE ( WHEN SUM(CASE WHEN TypeId ='10' THEN 1 ELSE 0 END) > 0 THEN 1 ELSE 0 END) +
CASE ( WHEN SUM(CASE WHEN TypeId ='5' THEN 1 ELSE 0 END) > 0 THEN 1 ELSE 0 END) +
CASE ( WHEN SUM(CASE WHEN TypeId ='8' THEN 1 ELSE 0 END) > 0 THEN 1 ELSE 0 END)
as SumOfNonZeros
FROM [Party]
Or maybe simpler
SELECT COUNT(*) AS Total,
COUNT(CASE WHEN TypeId ='4' THEN 1 END) AS 'TotalCount1',
COUNT(CASE WHEN TypeId ='6' THEN 1 END) AS 'TotalCount2',
COUNT(CASE WHEN TypeId ='1' THEN 1 END) AS 'TotalCount3',
COUNT(CASE WHEN TypeId ='10' THEN 1 END) AS 'TotalCount4',
COUNT(CASE WHEN TypeId ='5' THEN 1 END) AS 'TotalCount5',
COUNT(CASE WHEN TypeId ='8' THEN 1 END) AS 'TotalCount6',
COUNT( DISTINCT CASE WHEN TypeId IN ('4', '6', '1', '10', '5', '8')
THEN TypeId
END ) as CountOfNonZeros
FROM [Party]
Using CASE expression over the selecting result, you can do this:
SELECT (
CASE WHEN TotalCount1 > 0 THEN 1 ELSE 0 END +
CASE WHEN TotalCount2 > 0 THEN 1 ELSE 0 END +
CASE WHEN TotalCount3 > 0 THEN 1 ELSE 0 END +
CASE WHEN TotalCount4 > 0 THEN 1 ELSE 0 END +
CASE WHEN TotalCount5 > 0 THEN 1 ELSE 0 END +
CASE WHEN TotalCount6 > 0 THEN 1 ELSE 0 END) AS Result
FROM (
SELECT COUNT(*) AS Total,
SUM(CASE WHEN TypeId ='4' THEN 1 ELSE 0 END) AS 'TotalCount1',
SUM(CASE WHEN TypeId ='6' THEN 1 ELSE 0 END) AS 'TotalCount2',
SUM(CASE WHEN TypeId ='1' THEN 1 ELSE 0 END) AS 'TotalCount3',
SUM(CASE WHEN TypeId ='10' THEN 1 ELSE 0 END) AS 'TotalCount4',
SUM(CASE WHEN TypeId ='5' THEN 1 ELSE 0 END) AS 'TotalCount5',
SUM(CASE WHEN TypeId ='8' THEN 1 ELSE 0 END) AS 'TotalCount6'
FROM [Party]
) A

Trying to introduce a simple new condition to a JOIN query

I am using SQLite and I had someone help me construct this JOIN query which works quite well, but now I need to add another condition but I am having trouble introducing it to the query without it breaking.
In both tables used in the JOIN there is a column called EventId and I want to introduce the simple condition...
WHERE EventId = 123456
Below you can see a working example of the query itself along with two comments where I have tried to introduce the new condition and failed (because I'm bad at SQL).
SELECT t.MicrosoftId,
SUM(CASE WHEN name = 'necktie' THEN 1 ELSE 0 END) as 'necktie',
SUM(CASE WHEN name = 'shirt' THEN 1 ELSE 0 END) as 'shirt',
SUM(CASE WHEN name = 'suit' THEN 1 ELSE 0 END) as 'suit',
SUM(CASE WHEN name = 'man' THEN 1 ELSE 0 END) as 'man',
SUM(CASE WHEN name = 'male' THEN 1 ELSE 0 END) as 'male'
FROM TagsMSCV t
/* <---- WHERE t.EventId = 123456 (fails here...) */
LEFT JOIN
(SELECT i.MicrosoftId
FROM Images i
GROUP BY i.MicrosoftId) i
ON i.MicrosoftId = t.MicrosoftId
WHERE t.name IN ('necktie','shirt','suit','man','male')
/* <---- AND WHERE t.EventId = 123456 (fails here too...) */
GROUP BY t.MicrosoftId
try like below
select t1.* from ( SELECT t.MicrosoftId,
SUM(CASE WHEN name = 'necktie' THEN 1 ELSE 0 END) as 'necktie',
SUM(CASE WHEN name = 'shirt' THEN 1 ELSE 0 END) as 'shirt',
SUM(CASE WHEN name = 'suit' THEN 1 ELSE 0 END) as 'suit',
SUM(CASE WHEN name = 'man' THEN 1 ELSE 0 END) as 'man',
SUM(CASE WHEN name = 'male' THEN 1 ELSE 0 END) as 'male'
FROM TagsMSCV t WHERE t.EventId = 123456
and name IN ('necktie','shirt','suit','man','male') group by t.MicrosoftId
) t1
You did mistake to create subquery and as 2nd subquery no need group by as there no aggregate function used
It should be in WHERE section, but without second WHERE keyword:
SELECT t.MicrosoftId,
SUM(CASE WHEN name = 'necktie' THEN 1 ELSE 0 END) as 'necktie',
SUM(CASE WHEN name = 'shirt' THEN 1 ELSE 0 END) as 'shirt',
SUM(CASE WHEN name = 'suit' THEN 1 ELSE 0 END) as 'suit',
SUM(CASE WHEN name = 'man' THEN 1 ELSE 0 END) as 'man',
SUM(CASE WHEN name = 'male' THEN 1 ELSE 0 END) as 'male'
FROM TagsMSCV t
LEFT JOIN
(SELECT i.MicrosoftId
FROM Images i
GROUP BY i.MicrosoftId) i
ON i.MicrosoftId = t.MicrosoftId
WHERE t.name IN ('necktie','shirt','suit','man','male')
AND t.EventId = 123456
GROUP BY t.MicrosoftId

Finding the values from same table that have value 1 and 2 but not 3

Hi I have the below table where I want to find the object_ida that I have format_id 1,3,11 and 12 but not 10. Could you please help?
You can use group by and having:
select object_id
from t
group by object_id
having sum(case when format_id = 1 then 1 else 0 end) > 0 and
sum(case when format_id = 3 then 1 else 0 end) > 0 and
sum(case when format_id = 11 then 1 else 0 end) > 0 and
sum(case when format_id = 12 then 1 else 0 end) > 0 and
sum(case when format_id = 10 then 1 else 0 end) = 0;
Each condition tests for one of the format_ids. The > 0 means that at least one is assigned to an object_id. The = 0 means that none are assigned.

How do you create a conditional count across multiple fields?

I have the huge table of over 100 million rows of data which is joined to another reference table that I want to create a conditional count for.
The first table is the large one which is an audit log and contains data which lists data on countries and contains a date of audit.
The second table is a smaller table which contains relational data to the audit log.
The first part is the easy bit which is to identify which audit data I want to see. I have the following code to identify this:
select aud.*
from audit_log aud
join database db on db.id=aud.release_id
where aud.event_description like '% opted in'
and r.creation_source = 'system_a'
This gives me the data in the following format:
Country Event Description Audit Date
Czech Republic Czech Republic has been automatically opted in 11-AUG-14 07.01.52.606000000
Denmark Denmark has been automatically opted in 12-AUG-15 07.01.53.239000000
Denmark Denmark has been automatically opted in 11-SEP-15 07.01.53.902000000
Dominican Republic Dominican Republic has been automatically opted in 11-SEP-15 07.01.54.187000000
Ecuador Ecuador has been automatically opted in 11-DEC-14 07.01.54.427000000
Ecuador Ecuador has been automatically opted in 11-NOV-14 07.01.54.679000000
The number of results from this query still returns over 5 million rows so I cannot export the data to Excel to create a count.
My two main issues are the number of rows and the date format of the 'Audit Date' field.
Ideally I want to create a count which shows the data as:
Country |Aug-14|Nov-14|Dec-14|Aug-15|Sep-15
Czech Republic | 1 | | | |
Denmark | | | | 1 | 1
Dominican Republic | | | | | 1
Ecuador | | 1 | 1 | |
Any idea's on how I extract the month and year and drop the figures into column by country?
Thanks
Edit - Thank you xQbert for you solution, it worked perfectly!
The problem now is that I have run into a new problem.
I need to constrain the count by another query, but there is no unique identifier between the tables involved.
For example, I amended your query to fit my db:
select cty.country_name,
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='AUG-2014' then 1 else 0 end) as "AUG-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='SEP-2014' then 1 else 0 end) as "SEP-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='OCT-2014' then 1 else 0 end) as "OCT-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='NOV-2014' then 1 else 0 end) as "NOV-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='DEC-2014' then 1 else 0 end) as "DEC-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='JAN-2015' then 1 else 0 end) as "JAN-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='FEB-2015' then 1 else 0 end) as "FEB-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='MAR-2015' then 1 else 0 end) as "MAR-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='APR-2015' then 1 else 0 end) as "APR-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='MAY-2015' then 1 else 0 end) as "MAY-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='JUN-2015' then 1 else 0 end) as "JUN-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='JUL-2015' then 1 else 0 end) as "JUL-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='AUG-2015' then 1 else 0 end) as "AUG-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='SEP-2015' then 1 else 0 end) as "SEP-15"
from dschd.audit_trail aud
join dschd.release r on r.id=aud.release_id
join dschd.country cty on aud.EVENT_COUNTRY_ID=cty.id
where aud.event_description like '% opted in'
and r.creation_source = 'DSCHED'
GROUP BY cty.COUNTRY_name
My second query is:
select *
from DSCHD.RELEASE_COUNTRY_RIGHT rcr
join dschd.release r on rcr.RELEASE_ID=r.ID
join dschd.country cty on rcr.COUNTRY_ID=cty.id
where r.release_status in ('DRAFT', 'SCHEDULED', 'FINAL', 'DELIVERED')
and r.is_active = 'Y'
and rcr.MARKETING_RIGHT = 'Y'
and rcr.OPT_OUT = 'N'
and r.creation_source = 'DSCHED'
The problem is that I have many countries which can relate to one ID (Release_ID) but there is no unique identifier between the tables on a country level. Each country has an ID though.
So for query 1, to identify each unique row I would need the 'aud.Release_ID' and the 'aud.Event_country_id' and for query 2 to achieve the same I would need to use the 'rcr.Release_ID' and 'rcr.country_id'.
select cty.country_name,
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='AUG-2014' then 1 else 0 end) as "AUG-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='SEP-2014' then 1 else 0 end) as "SEP-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='OCT-2014' then 1 else 0 end) as "OCT-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='NOV-2014' then 1 else 0 end) as "NOV-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='DEC-2014' then 1 else 0 end) as "DEC-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='JAN-2015' then 1 else 0 end) as "JAN-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='FEB-2015' then 1 else 0 end) as "FEB-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='MAR-2015' then 1 else 0 end) as "MAR-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='APR-2015' then 1 else 0 end) as "APR-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='MAY-2015' then 1 else 0 end) as "MAY-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='JUN-2015' then 1 else 0 end) as "JUN-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='JUL-2015' then 1 else 0 end) as "JUL-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='AUG-2015' then 1 else 0 end) as "AUG-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='SEP-2015' then 1 else 0 end) as "SEP-15"
from dschd.audit_trail aud
join dschd.release r on r.id=aud.release_id
join dschd.country cty on aud.EVENT_COUNTRY_ID=cty.id
where aud.event_description like '% opted in'
and ***** in (select ******
from DSCHD.RELEASE_COUNTRY_RIGHT rcr
join dschd.release r on rcr.RELEASE_ID=r.ID
join dschd.country cty on rcr.COUNTRY_ID=cty.id
where r.release_status in ('DRAFT', 'SCHEDULED', 'FINAL', 'DELIVERED')
and r.is_active = 'Y'
and rcr.MARKETING_RIGHT = 'Y'
and rcr.OPT_OUT = 'N'
and r.creation_source = 'DSCHED')
GROUP BY cty.COUNTRY_name
The bit I am stuck at are the two parts which are indicated by '*****' as the join criteria is two fields.
Any ideas?
Quick and dirty, not dynamic floating based on a 12 month cylce or anything...
select country,
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='AUG-2014' then 1 else 0 end) as "AUG-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='SEP-2014' then 1 else 0 end) as "SEP-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='OCT-2014' then 1 else 0 end) as "OCT-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='NOV-2014' then 1 else 0 end) as "NOV-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='DEC-2014' then 1 else 0 end) as "DEC-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='JAN-2015' then 1 else 0 end) as "JAN-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='FEB-2015' then 1 else 0 end) as "FEB-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='MAR-2015' then 1 else 0 end) as "MAR-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='APR-2015' then 1 else 0 end) as "APR-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='MAY-2015' then 1 else 0 end) as "MAY-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='JUN-2015' then 1 else 0 end) as "JUN-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='JUL-2015' then 1 else 0 end) as "JUL-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='AUG-2015' then 1 else 0 end) as "AUG-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='SEP-2015' then 1 else 0 end) as "SEP-15"
from audit_log aud
join database db on db.id=aud.release_id
where aud.event_description like '% opted in'
and r.creation_source = 'system_a'
GROUP BY COUNTRY
Ideally we'd simply use a Pivot statement or base it on earliest date in range and go on... Such as found in this prior stack article Dynamic pivot in oracle sql
update based on changing requirements you do know you can join on multiple criteria right? :P
Note we created an inline view with your second query alias it as z table name and then add the two columns desired to match on as part of the results. Then we join it as if it were a table!
select cty.country_name,
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='AUG-2014' then 1 else 0 end) as "AUG-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='SEP-2014' then 1 else 0 end) as "SEP-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='OCT-2014' then 1 else 0 end) as "OCT-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='NOV-2014' then 1 else 0 end) as "NOV-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='DEC-2014' then 1 else 0 end) as "DEC-14",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='JAN-2015' then 1 else 0 end) as "JAN-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='FEB-2015' then 1 else 0 end) as "FEB-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='MAR-2015' then 1 else 0 end) as "MAR-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='APR-2015' then 1 else 0 end) as "APR-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='MAY-2015' then 1 else 0 end) as "MAY-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='JUN-2015' then 1 else 0 end) as "JUN-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='JUL-2015' then 1 else 0 end) as "JUL-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='AUG-2015' then 1 else 0 end) as "AUG-15",
SUM(CASE WHEN to_char(Audit_Date,'MON-YYYY') ='SEP-2015' then 1 else 0 end) as "SEP-15"
from dschd.audit_trail aud
join dschd.release r on r.id=aud.release_id
join dschd.country cty on aud.EVENT_COUNTRY_ID=cty.id
join (select Release_ID, country_id
from DSCHD.RELEASE_COUNTRY_RIGHT rcr
join dschd.release r on rcr.RELEASE_ID=r.ID
join dschd.country cty on rcr.COUNTRY_ID=cty.id
where r.release_status in ('DRAFT', 'SCHEDULED', 'FINAL', 'DELIVERED')
and r.is_active = 'Y'
and rcr.MARKETING_RIGHT = 'Y'
and rcr.OPT_OUT = 'N'
and r.creation_source = 'DSCHED') Z
ON aud.Release_ID = z.Realease_ID and
aud.Event_country_id = z.country_id
where aud.event_description like '% opted in'
GROUP BY cty.COUNTRY_name

How to select on table and count occurrences some values

I am asking you for a help because I do not know SQL very well.
I need to count occurrences some values from table column to achieve effect like statistics table which will look like at the picture below:
Needed result:
Comment:
My Result Table needs to have first two columns (contry and site) comes from first table "Violations" and next 5 columns which will contain numbers (count) of occurrences status_id in "Violations" in each of possible values of id from Status table.
Explanation:
So, I have existing two tables: Violations and Status. Please look at my sqlfiddle
Violations:
id long,
country varchar(20),
site varchar(20),
status_id long, <-- this is the id of status in Status table.
... other columns not important in this case
Status:
id long,
status long
Column "status" have values (1-4) which are mapped to string values: Suspected Violation (1), Confirmed Violation (2), Confirmed No Violation (3), Not Determined (4)
In result of my join (or based only on one table Violations) is to have table which should contain columns:
from Violations table: "Country" and "Site"
from Status table: "Suspected Violation", "Confirmed Violation", "Confirmed No Violation", "Not Determined", "Total" (where this columns are counters of occurrences in Violation table).
Current Status and new Requirements:
First try is done (thanks to bluefeet) below and is almost perfect...
select v.country,
v.site,
SUM(case when s.id = 1 then 1 else 0 end) Total_SuspectedViolations,
SUM(case when s.id = 2 then 1 else 0 end) Total_ConfirmedViolations,
SUM(case when s.id = 3 then 1 else 0 end) Total_ConfirmedNoViolations,
SUM(case when s.id = 4 then 1 else 0 end) Total_NotDetermined,
COUNT(*) Total
from violations v
inner join status s
on v.status_id = s.id
group by v.country, v.site
or without JOIN:
select v.country,
v.site,
SUM(case when v.status_id = 1 then 1 else 0 end) Total_SuspectedViolations,
SUM(case when v.status_id = 2 then 1 else 0 end) Total_ConfirmedViolations,
SUM(case when v.status_id = 3 then 1 else 0 end) Total_ConfirmedNoViolations,
SUM(case when v.status_id = 4 then 1 else 0 end) Total_NotDetermined,
COUNT(*) Total
from violations v
group by v.country, v.site
...but is not including 3 issues as you can see in the picture which should be. I mean:
"- All -" which should count occurrences for all countries
"- Unknown -" which should count occurrences for some not recognized countries
"- All -" (regarding to each country) - which should count occurrences within one country
Additional Explanation:
-Unknown- meaning:
Unknown should count occurrences for countries which for example do not exists in DB Country table or have a wrong name/id and that's why is treated here as Unknown (I forgot mention that there is table Country in DB).
The same for sites, Unknown for sites means that someone put wrong value in Violations.status_id not from range (1-4) because these are only acceptable values existing in Status table.
We can assume that table Country looks like:
Country:
id long,
name varchar(30)
Please help me to write correct sql query which would include these 3 conditions, because I have a big problem to do that.
The All case can be easily done using UNION statement (see sqlFiddle for results):
(SELECT v.country,
v.site,
SUM(CASE WHEN v.status_id = 1 THEN 1 ELSE 0 END) Total_SuspectedViolations,
SUM(CASE WHEN v.status_id = 2 THEN 1 ELSE 0 END) Total_ConfirmedViolations,
SUM(CASE WHEN v.status_id = 3 THEN 1 ELSE 0 END) Total_ConfirmedNoViolations,
SUM(CASE WHEN v.status_id = 4 THEN 1 ELSE 0 END) Total_NotDetermined,
COUNT(*) Total,
0 'isAll'
FROM violations v
GROUP BY v.country, v.site)
union(
SELECT v.country,
'- All -',
SUM(CASE WHEN v.status_id = 1 THEN 1 ELSE 0 END) Total_SuspectedViolations,
SUM(CASE WHEN v.status_id = 2 THEN 1 ELSE 0 END) Total_ConfirmedViolations,
SUM(CASE WHEN v.status_id = 3 THEN 1 ELSE 0 END) Total_ConfirmedNoViolations,
SUM(CASE WHEN v.status_id = 4 THEN 1 ELSE 0 END) Total_NotDetermined,
COUNT(*) Total,
1 'isAll'
FROM violations v
GROUP BY v.country)
UNION (
SELECT '- All -',
'- All -',
SUM(CASE WHEN v.status_id = 1 THEN 1 ELSE 0 END) Total_SuspectedViolations,
SUM(CASE WHEN v.status_id = 2 THEN 1 ELSE 0 END) Total_ConfirmedViolations,
SUM(CASE WHEN v.status_id = 3 THEN 1 ELSE 0 END) Total_ConfirmedNoViolations,
SUM(CASE WHEN v.status_id = 4 THEN 1 ELSE 0 END) Total_NotDetermined,
COUNT(*) Total,
1 'isAll'
FROM violations v)
ORDER BY country, isAll DESC, site
However, the performance may not be really great with that kind of query, so I'm not saying it's the best possible solution - but it works.
Version with 'Unknow'
http://www.sqlfiddle.com/#!2/abfb7/21
(SELECT IF(c.name IS NULL, '- Unknow -', c.name) as name,
v.site,
SUM(CASE WHEN v.status_id = 1 THEN 1 ELSE 0 END) Total_SuspectedViolations,
SUM(CASE WHEN v.status_id = 2 THEN 1 ELSE 0 END) Total_ConfirmedViolations,
SUM(CASE WHEN v.status_id = 3 THEN 1 ELSE 0 END) Total_ConfirmedNoViolations,
SUM(CASE WHEN v.status_id = 4 THEN 1 ELSE 0 END) Total_NotDetermined,
COUNT(*) Total,
0 'isAll'
FROM violations v LEFT JOIN country c ON c.name = v.country
GROUP BY c.name, v.site)
union(
SELECT IF(c.name IS NULL, '- Unknow -', c.name) as name,
'- All -',
SUM(CASE WHEN v.status_id = 1 THEN 1 ELSE 0 END) Total_SuspectedViolations,
SUM(CASE WHEN v.status_id = 2 THEN 1 ELSE 0 END) Total_ConfirmedViolations,
SUM(CASE WHEN v.status_id = 3 THEN 1 ELSE 0 END) Total_ConfirmedNoViolations,
SUM(CASE WHEN v.status_id = 4 THEN 1 ELSE 0 END) Total_NotDetermined,
COUNT(*) Total,
1 'isAll'
FROM violations v LEFT JOIN country c ON c.name = v.country
GROUP BY c.name)
UNION (
SELECT '- All -',
'- All -',
SUM(CASE WHEN v.status_id = 1 THEN 1 ELSE 0 END) Total_SuspectedViolations,
SUM(CASE WHEN v.status_id = 2 THEN 1 ELSE 0 END) Total_ConfirmedViolations,
SUM(CASE WHEN v.status_id = 3 THEN 1 ELSE 0 END) Total_ConfirmedNoViolations,
SUM(CASE WHEN v.status_id = 4 THEN 1 ELSE 0 END) Total_NotDetermined,
COUNT(*) Total,
1 'isAll'
FROM violations v LEFT JOIN country c ON c.name = v.country)
ORDER BY name, isAll DESC, site
Using with rollup you will get the needed sum field which you want
select v.country,
v.site,
SUM(case when s.id = 1 then 1 else 0 end) Total_SuspectedViolations,
SUM(case when s.id = 2 then 1 else 0 end) Total_ConfirmedViolations,
SUM(case when s.id = 3 then 1 else 0 end) Total_ConfirmedNoViolations,
SUM(case when s.id = 4 then 1 else 0 end) Total_NotDetermined,
COUNT(*) Total
from violations v
inner join status s
on v.status_id = s.id
group by v.country, v.site WITH ROLLUP
Hope this helps
REFER for with rollup documentation
FIDDLE