Sql server 为什么对更少行的全文查询比对更多行的同一查询慢？

此查询需要21秒：

SELECT Top 50 ROW_NUMBER() OVER (ORDER BY aDate DESC) AS Row, aID, aHeadline, aAuthor, aDate, TopStory, SiteSectionID, PrintSection, aSummary, small_image, aBody FROM articles 
WHERE aLive = -1 AND full_story = -1 AND aDate >= '9/3/2012' AND aDate <= '10/3/2012' AND (Contains((aAuthor,aBody,aHeadline,aSubhead,aSummary), '"kodak"')) 
ORDER BY aDate DESC

选择前50行编号（）（按aDate DESC排序）作为文章中的行、辅助、标题、作者、aDate、TopStory、SiteSectionID、PrintSection、aSummary、small\u image、aBody
其中aLive=-1和full_story=-1，aDate>='9/3/2012'，aDate='9/3/2002'，aDate检查执行Plant典型的优化器问题。基数估计错误。如果运行第一个查询，然后运行第二个查询，然后再次运行第一个查询，它是否仍在21秒内运行？我想知道第一个查询是否很慢，因为它是第一次运行并且正在生成执行计划，所以第二个查询正在重用该计划，因此运行得更快。
SELECT Top 50 ROW_NUMBER() OVER (ORDER BY aDate DESC) AS Row, aID, aHeadline, aAuthor, aDate, TopStory, SiteSectionID, PrintSection, aSummary, small_image, aBody FROM articles 
WHERE aLive = -1 AND full_story = -1 AND aDate >= '9/3/2002' AND aDate <= '10/3/2012' AND (Contains((aAuthor,aBody,aHeadline,aSubhead,aSummary), '"kodak"')) 
ORDER BY aDate DESC