I tried finding the answer, but maybe I am too new to MSSQL, I come from MySQL, so this is my question super simplified to go straight to the point:
Imagine we have a table "Things"
Thingie | Value
--------+-------
Thing1 | 10
Thing1 | 15
Thing1 | 16
In MySQL I could do something like this in a query:
SET #halfvalue := 0;
SELECT Thingie, Value,
(#halfvalue := Value / 2) AS HalfValue,
(#halfvalue / 2) AS HalfOfHalf
FROM Things
Which would return
Thingie | Value | HalfValue | HalfofHalf
--------+-------+-----------+------------
Thing1 | 10 | 5.00 | 2.50
Thing1 | 15 | 7.50 | 3.75
Thing1 | 16 | 8.00 | 4.00
Looks pretty simple, the actual one is a tad more complicated.
My problem is, in MSSQL I can't assign, and use a variable on the same SELECT. And I can't find anything similar to this functionality on this simple level.
Any solutions?
Edit, this is the select that contains all those nasty operations:
SELECT
fvh.DocEntry,
MAX( fvs.SeriesName ) AS "Serie",
MAX( fvh.DocNum - 1000000 ) AS "Número",
MAX( fvh.DocDate ) AS "Fecha",
MAX( fvh.U_FacNit ) AS "NIT",
MAX( fvh.U_FacNom ) AS "Nombre",
MAX( IIF( ISNULL( fvh.Address, '' ) = '', fvh.Address2, fvh.Address ) ) AS "Dirección",
SUM( fvd.GTotal - ISNULL( ncd.GTotal, 0 ) ) AS "Total",
IIF( MAX( fvh.CANCELED ) = 'Y' OR ( SUM( fvd.GTotal - ISNULL( ncd.GTotal, 0 ) ) = 0 ),
'Anulada',
IIF( SUM( fvd.GTotal ) > SUM( ISNULL( ncd.GTotal, 0 ) ) AND ( SUM( ISNULL( ncd.GTotal, 0 ) ) > 0 ),
'Devuelta',
'Emitida' )
) AS "Estado",
ROUND( ( ( SUM( fvd.GTotal - ISNULL( ncd.GTotal, 0 ) ) / 1.12 ) * 0.12 ), 4 ) AS "IVA",
ROUND( SUM( IIF( fvd.U_TipoA = 'BB',
( fvd.GTotal - ISNULL( ncd.GTotal, 0 ) ) - ( ( ( fvd.GTotal - ISNULL( ncd.GTotal, 0 ) ) / 1.12 ) * 0.12 ),
0 ) ), 4) AS "Bien",
ROUND( SUM( IIF( fvd.U_TipoA = 'S',
( fvd.GTotal - ISNULL( ncd.GTotal, 0 ) ) - ( ( ( fvd.GTotal - ISNULL( ncd.GTotal, 0 ) ) / 1.12 ) * 0.12 ),
0 ) ), 4) AS "Servicio",
ROUND( SUM( IIF( fvd.U_TipoA = 'N',
( fvd.GTotal - ISNULL( ncd.GTotal, 0 ) ) - ( ( ( fvd.GTotal - ISNULL( ncd.GTotal, 0 ) ) / 1.12 ) * 0.12 ),
0 ) ), 4) AS "No Aplica",
COUNT(fvd.LineNum) AS "Lineas", SUM(fvd.GTotal) AS "FCTotal",
SUM(ISNULL( ncd.GTotal, 0 )) AS "NCTotal"
/* Facturas */
FROM OINV AS fvh
LEFT JOIN NNM1 AS fvs ON fvs.Series = fvh.Series
LEFT JOIN INV1 as fvd ON fvd.DocEntry = fvh.DocEntry
/* Notas de Credito */
LEFT JOIN RIN1 AS ncd ON ncd.BaseEntry = fvh.DocEntry AND ncd.LineNum = fvd.LineNum
WHERE fvh.DocDate BETWEEN ? AND ? /*AND fvh.DocEntry = 1108*/
GROUP BY fvh.DocEntry
Thank you all for your time. I will dismantle my query and re-do it taking into consideration all of your input. Gracias, totales.
You think you can do this in MySQL:
SET #halfvalue := 0;
SELECT Thingie, Value,
(#halfvalue := Value / 2) AS HalfValue,
(#halfvalue / 2) AS HalfOfHalf
FROM Things;
But you are wrong. Why? MySQL -- as with every other database -- does not guarantee the order of evaluation of expression in a SELECT. The documentation even warns about this:
In the following statement, you might think that MySQL will evaluate #a first and then do an assignment second:
SELECT #a, #a:=#a+1, ...;
However, the order of evaluation for expressions involving user variables is undefined.
In both databases, you can use a subquery. In the most recent versions of MySQL (and just about any other database), you can also use a CTE:
SELECT Thingie, Value, HalfValue,
(HalfValue / 2) AS HalfOfHalf
FROM (SELECT t.*, (Value / 2) AS HalfValue
FROM Things t
) t;
The answer is simple: you can't do that in MSSQL, because when you try it you'll get:
Msg 141, Level 15, State 1, Line 3
A SELECT statement that assigns a value to a variable must not be combined with data-retrieval operations.
which you most probably experienced.
The most simple workaround would be:
SELECT Thingie, Value, Value/2, Value/4 from Things
Other method:
select Thingie, Value, HalfValue, HalfValue / 2 from (
SELECT Thingie, Value, Value / 2 HalfValue from Things
) a
No, that doesn't work in SQL. The parameter value is not set until the query completes. You can do it in two steps:
DECLARE #halfvalue FLOAT = 0;
SELECT #halfvalue = ([Value] / 2)
FROM Things ;
SELECT Thingie
, [Value]
, HalfValue = [Value]/2
, HalfAgainValue = #halfvalue / 2
FROM Things ;
Related
I'm pretty inexperienced with SQL in general, so I'm struggling with this. I know this looks messy and inefficient (so open to suggestions on improving that too!)
I have two queries that are pulling and aggregating data from three different tables. The first query is pulling from a single table, the second is aggregating data from 2 different tables.
SELECT
ifnull( `rt_poes`.`Account`, 'Total Compliant Spend' ) AS `Account`,
round( sum( IF ((( `rt_poes`.`Platform` = 'Google' ) AND ( `rt_poes`.`Ad_Type` = 'Search' )), `rt_poes`.`Cost`, 0 )), 2 ) AS `Compliant Google Search`,
round( sum( IF ((( `rt_poes`.`Platform` = 'Microsoft' ) AND ( `rt_poes`.`Ad_Type` = 'Search' )), `rt_poes`.`Cost`, 0 )), 2 ) AS `Compliant Bing Search`,
round( sum( IF ((( `rt_poes`.`Platform` = 'Google' ) AND ( `rt_poes`.`Ad_Type` = 'Shopping' )), `rt_poes`.`Cost`, 0 )), 2 ) AS `Compliant Google Shopping`,
round( sum( IF ((( `rt_poes`.`Platform` = 'Microsoft' ) AND ( `rt_poes`.`Ad_Type` = 'Shopping' )), `rt_poes`.`Cost`, 0 )), 2 ) AS `Compliant Bing Shopping`,
round( sum( IF ((( `rt_poes`.`Platform` = 'Google' ) AND ( `rt_poes`.`Ad_Type` = 'Search' )), `rt_poes`.`Cost`, 0 )), 2 ) +
round( sum( IF ((( `rt_poes`.`Platform` = 'Microsoft' ) AND ( `rt_poes`.`Ad_Type` = 'Search' )), `rt_poes`.`Cost`, 0 )), 2 )
AS `Compliant Country Spend`
FROM
`rt_poes`
WHERE
((
`rt_poes`.`field1` LIKE '% condition1%'
)
OR ( `rt_poes`.`field2` LIKE '% condition2%' )
GROUP BY
`rt_poes`.`Account` WITH ROLLUP
ORDER BY
sum( `rt_poes`.`Cost` )
This works well and generates a nice overview of the data I'm looking for as below.
The second query is:
SELECT ifnull( Account, 'Total All Platform Spend' ) AS Account, Round(SUM(Cost), 2) AS total_platform_costs
FROM
(SELECT Account, Cost FROM QTD_Account_Report
UNION ALL
SELECT AccountName, Spend FROM Bing_QTD_Account_Report) AS DerivedTable
GROUP BY DerivedTable.Account WITH ROLLUP
ORDER BY total_platform_costs
This generates another nice view of the data from another table that I'm looking for too.
The output of both is correct.
Now, I'm trying to join both on the account column. Essentially just add the second column from the second output on to the output of the first query.
The problem I'm having is that whenever I try to use a standard join, I get an error that MySQL does not support this syntax.
Is there another way to join the two or modify the second so that I can have a single view with both like below? Or do I have to create a whole new table for this?
I haven’t done MySQL in a while, but simply joining the second query should work. Is this what you did?
select
ifnull(rt_poes.Account, 'Total Compliant Spend') as Account,
round( sum( if(rt_poes.Platform = 'Google' AND rt_poes.Ad_Type = 'Search', rt_poes.Cost, 0)), 2) as `Compliant Google Search`,
round( sum( if(rt_poes.Platform = 'Microsoft' AND rt_poes.Ad_Type = 'Search', rt_poes.Cost, 0)), 2) as `Compliant Bing Search`,
round( sum( if(rt_poes.Platform = 'Google' AND rt_poes.Ad_Type = 'Shopping', rt_poes.Cost, 0)), 2) as `Compliant Google Shopping`,
round( sum( if(rt_poes.Platform = 'Microsoft' AND rt_poes.Ad_Type = 'Shopping', rt_poes.Cost, 0)), 2) as `Compliant Bing Shopping`,
round( sum( if(rt_poes.Platform = 'Google' AND rt_poes.Ad_Type = 'Search', rt_poes.Cost, 0)), 2) +
round( sum( if(rt_poes.Platform = 'Microsoft' AND rt_poes.Ad_Type = 'Search', rt_poes.Cost, 0)), 2) as `Compliant Country Spend`
round(SUM(DerivedTable.Cost), 2) as total_platform_costs
from rt_poes
left join (select Account, Cost
from QTD_Account_Report
UNION ALL
select AccountName, Spend
from Bing_QTD_Account_Report) as DerivedTable on DerivedTable.Account = rt_poes.Account
where rt_poes.field1 like '%condition1%' or rt_poes.field2 like '%condition2%'
group by rt_poes.Account WITH ROLLUP
order by sum(rt_poes.Cost)
It would be helpful if you quoted the specific error message you got, as well as the query that caused it.
I get error
SQL Error (1055): Expression #7 of SELECT list is not in GROUP BY clause and contains nonaggregated column 'ifu.amount' which is not functionally dependent on columns in GROUP BY clause; this is incompatible with sql_mode=only_full_group_by
after migrating to mysql 8.0 from 5.6. I know that it can be easily fixed by disabling ONLY_FULL_GROUP_BY flag, but I want it to be more compatible with mysql 8.0. So question is if I would add ifu.amount to GROUP BY it should work perfetcly fine and I won't miss any query results or anything? Now without GROUP BY ifu.amount MySQL code looks like:
select
`i`.`id` AS `institution_id`,
`i`.`name` AS `institution_name`,
`cr`.`check_date` AS `check_date`,
sum(
(
case when (`cr`.`status` = '1') then 1 else 0 end
)
) AS `can_accept`,
sum(
(
case when (`cr`.`status` = '0') then 1 else 0 end
)
) AS `cannot_accept`,(
sum(
(
case when (`cr`.`status` = '1') then 1 else 0 end
)
) + sum(
(
case when (`cr`.`status` = '0') then 1 else 0 end
)
)
) AS `suma`,
`ifu`.`amount` AS `amount`,
round(
(
(
(
(
sum(
(
case when (`cr`.`status` = '1') then 1 else 0 end
)
) * 100
) / (
sum(
(
case when (`cr`.`status` = '1') then 1 else 0 end
)
) + sum(
(
case when (`cr`.`status` = '0') then 1 else 0 end
)
)
)
) * `ifu`.`amount`
) * 0.01
),
2
) AS `financed_amount`
from
(
(
(
`check_results` `cr`
join `family_doctors` `fd` on((`fd`.`id` = `cr`.`doctor_id`))
)
join `institutions` `i` on((`i`.`id` = `fd`.`institution_id`))
)
join `institutions_funding` `ifu` on((`ifu`.`institution_id` = `i`.`id`))
)
where
(`cr`.`status` in (1, 0))
group by
`i`.`id`,
`i`.`name`,
`cr`.`check_date`
Thanks for help in advance!
Include amount in your group by clause.
where
(`cr`.`status` in (1, 0))
group by
`i`.`id`,
`i`.`name`,
`cr`.`check_date`,
`ifu`.`amount`
if amount is excluded on your group by clause, this will get the amount that corresponds on your id, name and check date in ascending order (default).
or
min(`ifu`.`amount`) as `amount`.
I have this crazy nested sql statement and it returns something like this:
However, what I want is to use this data in the image to return 2 columns. col_1:SUM(avgWithCriteria)/43 and col_2:SUM(avgWithoutCriteria)/43. How can I do this by adding onto my query below?
SELECT
( ( avgWithCriteria - totalAverage ) / ( ( avgWithCriteria + totalAverage ) / 2 ) ) * 100 as percentDifference,
a.*
FROM
(SELECT
AVG( CASE WHEN 'f' not in ( has_free_parking ) THEN price ELSE null END) as avgWithCriteria,
AVG( CASE WHEN 'f' in ( has_free_parking ) THEN price ELSE null END) as avgWithoutCriteria,
AVG( price ) as totalAverage,
neighbourhood_cleansed
FROM listings
WHERE city_name="berlin"
AND price <= 1000000
AND price >= -1
AND reviews_per_month <= 1000000
AND reviews_per_month >= -1
AND est_monthly_income <= 1000000
AND est_monthly_income >= -1
GROUP BY neighbourhood_cleansed ) a;
Try this:-
SELECT percentDifference,avgWithCriteria,avgWithoutCriteria,totalAverage,neighbourhood_cleansed,
(col1_1/43) as col1,(col2_2/43) as col2
from
(
SELECT a.*,SUM(avgWithCriteria) as col1_1,SUM(avgWithoutCriteria) as col2_2
FROM
(
SELECT
( ( avgWithCriteria - totalAverage ) / ( ( avgWithCriteria + totalAverage ) / 2 ) ) * 100 as percentDifference,
a.*
FROM
(SELECT
AVG( CASE WHEN 'f' not in ( has_free_parking ) THEN price ELSE null END) as avgWithCriteria,
AVG( CASE WHEN 'f' in ( has_free_parking ) THEN price ELSE null END) as avgWithoutCriteria,
AVG( price ) as totalAverage,
neighbourhood_cleansed
FROM listings
WHERE city_name="berlin"
AND price <= 1000000
AND price >= -1
AND reviews_per_month <= 1000000
AND reviews_per_month >= -1
AND est_monthly_income <= 1000000
AND est_monthly_income >= -1
GROUP BY neighbourhood_cleansed ) a
) a
GROUP BY percentDifference,avgWithCriteria,avgWithoutCriteria,totalAverage,neighbourhood_cleansed
) a;
SELECT
( ( avgWithCriteria - totalAverage ) / ( ( avgWithCriteria + totalAverage ) / 2 ) ) * 100 as percentDifference,
SUM(avgWithCriteria)/43 AS col_1,
SUM(avgWithoutCriteria)/43 AS col_2,
a.*
FROM
(SELECT
AVG( CASE WHEN 'f' not in ( has_free_parking ) THEN price ELSE null END) as avgWithCriteria,
AVG( CASE WHEN 'f' in ( has_free_parking ) THEN price ELSE null END) as avgWithoutCriteria,
AVG( price ) as totalAverage,
neighbourhood_cleansed
FROM listings
WHERE city_name="berlin"
AND price <= 1000000
AND price >= -1
AND reviews_per_month <= 1000000
AND reviews_per_month >= -1
AND est_monthly_income <= 1000000
AND est_monthly_income >= -1
GROUP BY neighbourhood_cleansed ) a;
I am trying to perform a group summation in mysql , but the nulls are causing issue. I used nullif/ifnull none helped
Here is an example
Group Employee Discipline HardWork Dedication
101 DTF 1 1 1
101 Tim 1 1 0
101 Erk NULL 1 0
101 PMD NULL 1 0
101 NSE 1 1 0
SQL:
SELECT
GROUP,
SUM( EMPLOYEE ),
SUM( DISCIPLINE ),
SUM( HARDWORK ),
SUM( DEDICATION )
FROM
TABLE
GROUP BY
GROUP,
EMPLOYEE,
DISCIPLINE,
HARDWORK,
DEDICATION
Expected result:
101---3----5----1
Actual result
101--5----5-----1
Anyone has any ideas?
Try:
SUM(IFNULL( Discipline, 0 ))
instead of
SUM( DISCIPLINE )
The firt one will take the value 0 is a null is found.
I think your query should be:
SELECT
Group,
COUNT( Employee ) AS EmployeeCountInGroup,
SUM( ISNULL( Discipline, 0 ) ) AS SumDiscipline,
SUM( ISNULL( HardWork, 0 ) ) AS SumHardWork,
SUM( ISNULL( Dedication, 0 ) ) AS SumDedication
FROM
TableName
GROUP BY
Group
I have a question and answers listing and an option to filter the questions based on the % of correct answers. So I am using the following query for the listing :
SELECT
question_id,
text
FROM
test_answers LEFT JOIN test_questions ON test_questions.id = test_answers.question_id
LEFT JOIN test_categories ON test_questions.`category_id` = test_categories.id
WHERE `question_id` IN(question IDS)
GROUP BY `question_id`
ORDER BY `question_id` DESC;
and using another query for finding the question IDS for which the % of correct answers in the given range. The query is as follows :
SELECT q1.question_id FROM (
SELECT test_answers.question_id AS question_id,
SUM( IF( test_answers.correct_answer =1, 1, 0 ) ) AS correct_answers,
SUM( IF( test_answers.correct_answer !=1, 1, 0 ) ) AS incorrect_answers,
round( ( SUM( IF( test_answers.correct_answer =1, 1, 0 ) ) / ( SUM( IF( test_answers.correct_answer =1, 1, 0 ) ) + SUM( IF( test_answers.correct_answer !=1, 1, 0 ) ) ) *100 ) , 2 ) AS percentage
FROM test_replies
JOIN test_answers ON test_replies.answer_id = test_answers.id
GROUP BY test_answers.question_id
HAVING percentage between 80 and 89 AND correct_answers >25
) AS q1
Now the issue is that the second query returns almost 4000 question Ids and it will increase in the near future and might be become 10k or more. So I seriously would like to optimize the query as it is going to impact the performance in a great deal. Can anyone suggest a better method for doing it ?
try join instead of IN, see if it helps. (sql not tested)
SELECT
ta.question_id, text
FROM
(
SELECT test_answers.question_id AS question_id,
SUM( IF( test_answers.correct_answer =1, 1, 0 ) ) AS correct_answers,
SUM( IF( test_answers.correct_answer !=1, 1, 0 ) ) AS incorrect_answers,
round( ( SUM( IF( test_answers.correct_answer =1, 1, 0 ) ) / ( SUM( IF( test_answers.correct_answer =1, 1, 0 ) ) + SUM( IF( test_answers.correct_answer !=1, 1, 0 ) ) ) *100 ) , 2 ) AS percentage
FROM test_replies
JOIN test_answers ON test_replies.answer_id = test_answers.id
GROUP BY test_answers.question_id
HAVING percentage between 80 and 89 AND correct_answers >25
) AS q1
INNER JOIN
test_answers ta USING (question_id)
LEFT JOIN
test_questions ON test_questions.id = ta.question_id
LEFT JOIN
test_categories ON test_questions.`category_id` = test_categories.id
GROUP BY
ta.question_id`
ORDER BY
ta.question_id DESC;