一部のデータのカウントを返すクエリがありますが、null 値を含むデータは必要ありません...
例として、コードは統計情報をクリック システムからテーブルにロールします。
SELECT sh.dropid,
COUNT(DISTINCT IF(sh2.`SentDate` IS NOT NULL, sh2.`subid`, NULL)) AS sentCount,
COUNT(DISTINCT IF(sh2.`Open` = 1, sh2.`subid`, NULL)) AS openCount,
COUNT(DISTINCT IF(sh2.`Click` = 1, sh2.`subhistid`, NULL)) AS clickCount,
COUNT(DISTINCT IF(sh2.`Bounced` = 1, sh2.`subid`, NULL)) AS bounceCount,
COUNT(DISTINCT IF(sh2.`Unsubscribed` = 1, sh2.`subid`, NULL)) AS unsubCount,
COUNT(DISTINCT IF(sh2.`Abuse` = 1, sh2.`subid`, NULL)) AS abuseCount
FROM subscriberhistory sh
INNER JOIN subscriberhistory sh2 ON sh.subid = sh2.subid
WHERE sh.SentDate >= '#runDate# #lastRunTime#'
AND sh.dropid IS NOT NULL
AND sh.dropid != ""
OR (sh.SentDate IS NULL AND sh.OpenDate >= '#runDate# #lastRunTime#')
OR (sh.SentDate IS NULL AND sh.ClickDate >= '#runDate# #lastRunTime#')
OR (sh.SentDate IS NULL AND sh.UnsubscribeDate >= '#runDate# #lastRunTime#')
OR (sh.SentDate IS NULL AND sh.BouncedDate >= '#runDate# #lastRunTime#')
OR (sh.SentDate IS NULL AND sh.AbuseDate >= '#runDate# #lastRunTime#')
GROUP BY dropid
ORDER BY sentCount DESC
編集: 修正されたクエリ
SELECT sh.dropid,
COUNT(DISTINCT IF(sh2.`SentDate` IS NOT NULL, sh2.`subid`, NULL)) AS sentCount,
COUNT(DISTINCT IF(sh2.`Open` = 1, sh2.`subid`, NULL)) AS openCount,
COUNT(DISTINCT IF(sh2.`Click` = 1, sh2.`subid`, NULL)) AS clickCount,
COUNT(DISTINCT IF(sh2.`Bounced` = 1, sh2.`subid`, NULL)) AS bounceCount,
COUNT(DISTINCT IF(sh2.`Unsubscribed` = 1, sh2.`subid`, NULL)) AS unsubCount,
COUNT(DISTINCT IF(sh2.`Abuse` = 1, sh2.`subid`, NULL)) AS abuseCount
FROM subscriberhistory sh
INNER JOIN subscriberhistory sh2 ON sh.subid = sh2.subid
WHERE sh.dropid IS NOT NULL AND sh.dropid != ""
AND ((sh.SentDate >= '2012-04-10 14:15')
OR (sh.SentDate IS NULL AND sh.OpenDate >= '2012-04-10 14:15')
OR (sh.SentDate IS NULL AND sh.ClickDate >= '2012-04-10 14:15')
OR (sh.SentDate IS NULL AND sh.UnsubscribeDate >= '2012-04-10 14:15')
OR (sh.SentDate IS NULL AND sh.BouncedDate >= '2012-04-10 14:15')
OR (sh.SentDate IS NULL AND sh.AbuseDate >= '2012-04-10 14:15'))
GROUP BY sh.dropid
ORDER BY sentCount DESC
返されるレコード セットの例は次のようになります...
(dropid) sent opens clicks
400 2 3 4
401 2 3 6
NULL 2 3 4
繰り返しますが、目標は、前のレコード セットから NULL データを除外することです。この動作が発生する理由と、それを解決するために何をすべきかを誰かに説明してもらえますか?