强类型数据集中子查询的计算

Question

以下背景：我有一个表值函数，它返回一个表，该表是大表上的一种过滤视图（视图不能参数化）。现在我需要不同条件下的许多行计数值。是否可以在一个查询中获取所有计数值，而不必为每个条件查询此函数？

为了完整起见（没关系，你可以阅读它），这是函数：

CREATE FUNCTION [dbo].[getRuleConfigBySetSlsRules] (
 @fiRuleSet AS TinyInt,
 @fiServiceLocations AS NVARCHAR(4000),
 @fiRules AS NVARCHAR(400)
)

RETURNS TABLE AS RETURN  (
 SELECT     ruleChangeTo.fiSet, ruleChangeTo.fiClaimStatus, ruleChangeTo.fiRule, ruleChangeTo.fiSL, ruleChangeTo.fiResult, 
                      dimClaimStatus.ClaimStatusName AS OldStatus, dimClaimStatus_1.ClaimStatusName AS NewStatus, locSL.SLName, ruleSet.SetName, 
                      ruleResult.ResultName, locMarketUnit.MarketUnitName, locCountry.CountryName, ruleRule.RuleKey, ruleRule.RuleName, ruleRule.RuleDescription, 
                      locGSP.WCMSName AS GSP_WCMSName, locGSP.Active AS GSP_Active, locGSP.WCMSKeyNumber AS GSP_WCMSKeyNumber, 
                      locSL.Active AS SL_Active, locSL.WCMSName AS SL_WCMSName, locSL.WCMSKeyNumber AS SL_WCMSKeyNumber, locSL.fiSLType, 
                      locSLType.SLTypeName, ruleChangeTo.fiClaimStatus_ChangeTo
FROM         locGSP INNER JOIN
                      locCountry ON locGSP.fiCountry = locCountry.idCountry INNER JOIN
                      locMarketUnit ON locCountry.fiMarketUnit = locMarketUnit.idMarketUnit INNER JOIN
                      locSL ON locGSP.idGSP = locSL.fiGSP INNER JOIN
                      ruleChangeTo ON locSL.idSL = ruleChangeTo.fiSL INNER JOIN
                      dimClaimStatus ON ruleChangeTo.fiClaimStatus = dimClaimStatus.idClaimStatus INNER JOIN
                      ruleResult ON ruleChangeTo.fiResult = ruleResult.idResult INNER JOIN
                      ruleRule ON ruleChangeTo.fiRule = ruleRule.idRule INNER JOIN
                      ruleSet ON ruleChangeTo.fiSet = ruleSet.idSet INNER JOIN
                      locSLType ON locSL.fiSLType = locSLType.idSLType INNER JOIN
                      dimClaimStatus AS dimClaimStatus_1 ON ruleChangeTo.fiClaimStatus_ChangeTo = dimClaimStatus_1.idClaimStatus  INNER JOIN
                       dbo.Split(@fiServiceLocations, ',') AS ServiceLocations ON ServiceLocations.Item = ruleChangeTo.fiSL INNER JOIN
                      dbo.Split(@fiRules, ',') AS Rules ON ruleChangeTo.fiRule = Rules.Item
WHERE (ruleChangeTo.fiSet = @fiRuleSet)
)

通过以下方式我可以获得一个计数值，但实际上我需要 30 个条件来处理同一个查询（WHERE 子句）。

在没有对每个值进行一次查询的情况下，在强类型数据集中完成此任务的最佳方法是什么？我不明白。 这是必须使用各种状态条件及其计数值扩展的查询（在上图中）：

SELECT     COUNT(*) AS CountValidReject
FROM         dbo.getRuleConfigBySetSlsRules(@fiSet, @ServiceLocations, @Rules) AS RuleConfigFilter
WHERE     (fiClaimStatus = 1) AND (fiClaimStatus_ChangeTo = 8)

言归正传：我需要一个带有临时表或类似东西的存储过程，还是我可以在数据集中的这个 sql 中创建和查询一个临时表？我不想为每个值调用 getRuleConfigBySetSlsRules。

谢谢

编辑： 这是最后一个查询，它返回 1 行，其中包含许多计数值作为列。再次感谢 Steve 的 CASE WHEN 提示。效率很高。

SELECT  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 5) AND (fiClaimStatus_ChangeTo = 2) THEN 1 ELSE NULL END) AS RuleAppliesApprovedInvalid,
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 5) AND (fiClaimStatus_ChangeTo = 8) THEN 1 ELSE NULL END) AS RuleAppliesApprovedReject,
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 5) AND (fiClaimStatus_ChangeTo = 7) THEN 1 ELSE NULL END) AS RuleAppliesApprovedReview, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 9) AND (fiClaimStatus_ChangeTo = 2) THEN 1 ELSE NULL END) AS RuleAppliesEscalatedInvalid, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 9) AND (fiClaimStatus_ChangeTo = 8) THEN 1 ELSE NULL END) AS RuleAppliesEscalatedReject, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 9) AND (fiClaimStatus_ChangeTo = 7) THEN 1 ELSE NULL END) AS RuleAppliesEscalatedReview, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 9) AND (fiClaimStatus_ChangeTo = 1) THEN 1 ELSE NULL END) AS RuleAppliesEscalatedValid, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 8) AND (fiClaimStatus_ChangeTo = 2) THEN 1 ELSE NULL END) AS RuleAppliesRejectInvalid, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 8) AND (fiClaimStatus_ChangeTo = 1) THEN 1 ELSE NULL END) AS RuleAppliesRejectValid, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 7) AND (fiClaimStatus_ChangeTo = 2) THEN 1 ELSE NULL END) AS RuleAppliesReviewInvalid, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 7) AND (fiClaimStatus_ChangeTo = 8) THEN 1 ELSE NULL END) AS RuleAppliesReviewReject, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 7) AND (fiClaimStatus_ChangeTo = 1) THEN 1 ELSE NULL END) AS RuleAppliesReviewValid, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 1) AND (fiClaimStatus_ChangeTo = 5) THEN 1 ELSE NULL END) AS RuleAppliesValidApproved, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 1) AND (fiClaimStatus_ChangeTo = 2) THEN 1 ELSE NULL END) AS RuleAppliesValidInvalid, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 1) AND (fiClaimStatus_ChangeTo = 8) THEN 1 ELSE NULL END) AS RuleAppliesValidReject, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 1) AND (fiClaimStatus_ChangeTo = 7) THEN 1 ELSE NULL END) AS RuleAppliesValidReview, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 5) AND (fiClaimStatus_ChangeTo = 2) THEN 1 ELSE NULL END) AS RuleApplNotApprovedInvalid, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 5) AND (fiClaimStatus_ChangeTo = 8) THEN 1 ELSE NULL END) AS RuleApplNotApprovedReject, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 5) AND (fiClaimStatus_ChangeTo = 7) THEN 1 ELSE NULL END) AS RuleApplNotApprovedReview, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 9) AND (fiClaimStatus_ChangeTo = 2) THEN 1 ELSE NULL END) AS RuleApplNotEscalatedInvalid, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 9) AND (fiClaimStatus_ChangeTo = 8) THEN 1 ELSE NULL END) AS RuleApplNotEscalatedReject, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 9) AND (fiClaimStatus_ChangeTo = 7) THEN 1 ELSE NULL END) AS RuleApplNotEscalatedReview, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 9) AND (fiClaimStatus_ChangeTo = 1) THEN 1 ELSE NULL END) AS RuleApplNotEscalatedValid, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 8) AND (fiClaimStatus_ChangeTo = 2) THEN 1 ELSE NULL END) AS RuleApplNotRejectInvalid, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 8) AND (fiClaimStatus_ChangeTo = 1) THEN 1 ELSE NULL END) AS RuleApplNotRejectValid, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 7) AND (fiClaimStatus_ChangeTo = 2) THEN 1 ELSE NULL END) AS RuleApplNotReviewInvalid, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 7) AND (fiClaimStatus_ChangeTo = 8) THEN 1 ELSE NULL END) AS RuleApplNotReviewReject, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 7) AND (fiClaimStatus_ChangeTo = 1) THEN 1 ELSE NULL END) AS RuleApplNotReviewValid, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 1) AND (fiClaimStatus_ChangeTo = 5) THEN 1 ELSE NULL END) AS RuleApplNotValidApproved, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 1) AND (fiClaimStatus_ChangeTo = 2) THEN 1 ELSE NULL END) AS RuleApplNotValidInvalid, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 1) AND (fiClaimStatus_ChangeTo = 8) THEN 1 ELSE NULL END) AS RuleApplNotValidReject, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 1) AND (fiClaimStatus_ChangeTo = 7) THEN 1 ELSE NULL END) AS RuleApplNotValidReview, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 1) THEN 1 ELSE NULL END) AS AllValidApplies, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 1) THEN 1 ELSE NULL END) AS AllValidNotApplies, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 5) THEN 1 ELSE NULL END) AS AllApprovedApplies,          
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 5) THEN 1 ELSE NULL END) AS AllApprovedNotApplies, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 7) THEN 1 ELSE NULL END) AS AllReviewApplies, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 7) THEN 1 ELSE NULL END) AS AllReviewNotApplies, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 7) THEN 1 ELSE NULL END) AS AllRejectApplies, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 7) THEN 1 ELSE NULL END) AS AllRejectNotApplies, 
  COUNT(CASE WHEN (fiResult = 1) AND (fiClaimStatus = 9) THEN 1 ELSE NULL END) AS AllEscalatedApplies, 
  COUNT(CASE WHEN (fiResult = 2) AND (fiClaimStatus = 9) THEN 1 ELSE NULL END) AS AllEscalatedNotApplies
FROM         dbo.getRuleConfigBySlsRules(@ServiceLocations, @Rules) AS RuleConfig
WHERE     (fiSet = @fiSet)

Answer 1

如果你能更具体一些并举个例子，那会很有帮助。我不确定“同一查询的 30 个条件（WHERE 子句）”是什么意思，或者“强类型”与任何事情有什么关系。你能举两个或三个你想要的 30 样东西的例子吗？

你想要的猜测是这样的：

SELECT
  COUNT(CASE WHEN (fiClaimStatus = 1) AND (fiClaimStatus_ChangeTo = 8)
        THEN 1 ELSE NULL END) AS myCount1,
  COUNT(CASE WHEN (fiClaimStatus = 2) AND (fiClaimStatus_ChangeTo = 8)
        THEN 1 ELSE NULL END) AS myCount2,
  COUNT(CASE WHEN (fiClaimStatus = 2) AND (fiClaimStatus_ChangeTo = 6)
        THEN 1 ELSE NULL END) AS myCount1,
  ...
FROM dbo.getRuleConfigBySetSlsRules(@fiSet, @ServiceLocations, @Rules) AS T

如果这不符合您的需要，请提供更多信息。

Answer 2

我看不出有什么办法。将表值函数的结果放入临时表中，然后从中运行聚合查询。