web-dev-qa-db-fra.com

Décompte conditionnel sur un champ

Si j'avais une table comme celle-ci:

jobId, jobName, Priority

La priorité peut être un entier compris entre 1 et 5.

Étant donné que j'aurais besoin de cette requête pour générer un graphique sur le rapport, j'aurais besoin d'afficher le jobid, le nom du travail et 5 champs appelés Priority1, Priority2, Priority3, Priority4. Priorité 5.

Priority1 doit compter le nombre de lignes où le champ prioritaire a la valeur 1.

Priority2 doit compter le nombre de lignes où le champ prioritaire a la valeur 2.

Priority3 doit compter le nombre de lignes où le champ prioritaire a la valeur 3.

etc

Comment pourrais-je faire cela d'une manière rapide et performante?

Merci beaucoup, Kave

sqlcountconditional
56
Houman

Je pense que tu peux être après

select 
    jobID, JobName,
    sum(case when Priority = 1 then 1 else 0 end) as priority1,
    sum(case when Priority = 2 then 1 else 0 end) as priority2,
    sum(case when Priority = 3 then 1 else 0 end) as priority3,
    sum(case when Priority = 4 then 1 else 0 end) as priority4,
    sum(case when Priority = 5 then 1 else 0 end) as priority5
from
    Jobs
group by 
    jobID, JobName

Cependant, je ne suis pas certain que vous ayez besoin de jobID et JobName dans vos résultats, le cas échéant, supprimez-les et supprimez le groupe par,

106
David Raznick

L'utilisation de COUNT au lieu de SUM supprime l'exigence d'une instruction ELSE:

SELECT jobId, jobName,
    COUNT(CASE WHEN Priority=1 THEN 1 END) AS Priority1,
    COUNT(CASE WHEN Priority=2 THEN 1 END) AS Priority2,
    COUNT(CASE WHEN Priority=3 THEN 1 END) AS Priority3,
    COUNT(CASE WHEN Priority=4 THEN 1 END) AS Priority4,
    COUNT(CASE WHEN Priority=5 THEN 1 END) AS Priority5
FROM TableName
GROUP BY jobId, jobName
71
taur

IIF n'est pas une construction SQL standard, mais si elle est prise en charge par votre base de données, vous pouvez obtenir une instruction plus élégante produisant le même résultat:

SELECT JobId, JobName,

COUNT(IIF (Priority=1, 1, NULL)) AS Priority1,
COUNT(IIF (Priority=2, 1, NULL)) AS Priority2,
COUNT(IIF (Priority=3, 1, NULL)) AS Priority3,
COUNT(IIF (Priority=4, 1, NULL)) AS Priority4,
COUNT(IIF (Priority=5, 1, NULL)) AS Priority5

FROM TableName
GROUP BY JobId, JobName
5
Guy Levin

En utilisant des instructions CASE ANSI SQL-92, vous pouvez faire quelque chose comme ceci (tableau dérivé plus cas):

 SELECT jobId, jobName, SUM(Priority1)
 AS Priority1, SUM(Priority2) AS
 Priority2, SUM(Priority3) AS
 Priority3, SUM(Priority4) AS
 Priority4,  SUM(Priority5) AS
 Priority5 FROM (
     SELECT jobId, jobName,
     CASE WHEN Priority = 1 THEN 1 ELSE 0 END AS Priority1,
     CASE WHEN Priority = 2 THEN 1 ELSE 0 END AS Priority2,
     CASE WHEN Priority = 3 THEN 1 ELSE 0 END AS Priority3,
     CASE WHEN Priority = 4 THEN 1 ELSE 0 END AS Priority4,
     CASE WHEN Priority = 5 THEN 1 ELSE 0 END AS Priority5
     FROM TableName


)
4
Anon246
SELECT  Priority, COALESCE(cnt, 0)
FROM    (
        SELECT  1 AS Priority
        UNION ALL
        SELECT  2 AS Priority
        UNION ALL
        SELECT  3 AS Priority
        UNION ALL
        SELECT  4 AS Priority
        UNION ALL
        SELECT  5 AS Priority
        ) p
LEFT JOIN
        (
        SELECT  Priority, COUNT(*) AS cnt
        FROM    jobs
        GROUP BY
                Priority
        ) j
ON      j.Priority = p.Priority
1
Quassnoi

Vous pouvez rejoindre la table contre elle-même:

select
   t.jobId, t.jobName,
   count(p1.jobId) as Priority1,
   count(p2.jobId) as Priority2,
   count(p3.jobId) as Priority3,
   count(p4.jobId) as Priority4,
   count(p5.jobId) as Priority5
from
   theTable t
   left join theTable p1 on p1.jobId = t.jobId and p1.jobName = t.jobName and p1.Priority = 1
   left join theTable p2 on p2.jobId = t.jobId and p2.jobName = t.jobName and p2.Priority = 2
   left join theTable p3 on p3.jobId = t.jobId and p3.jobName = t.jobName and p3.Priority = 3
   left join theTable p4 on p4.jobId = t.jobId and p4.jobName = t.jobName and p4.Priority = 4
   left join theTable p5 on p5.jobId = t.jobId and p5.jobName = t.jobName and p5.Priority = 5
group by
   t.jobId, t.jobName

Ou vous pouvez utiliser la casse dans une somme:

select
   jobId, jobName,
   sum(case Priority when 1 then 1 else 0 end) as Priority1,
   sum(case Priority when 2 then 1 else 0 end) as Priority2,
   sum(case Priority when 3 then 1 else 0 end) as Priority3,
   sum(case Priority when 4 then 1 else 0 end) as Priority4,
   sum(case Priority when 5 then 1 else 0 end) as Priority5
from
   theTable
group by
   jobId, jobName
1
Guffa