Left-duplicate indexes
At Kim Tripp's PASS session on indexing and covered indexes, she more-or-less proved that wider is better for indexes, within reason and searchability. It's better to have the index (Column1, Column2, Column3) than the index (Column1) and then a (Column1, Column2) and then a (Column1, Column2, Column3). And the first two of that set are redundant and don't provide any significant performance improvement over a wider index in most applications.Finding cases where you have multiple indexes like that isn't a trivial problem. Here's the answer I came up with. This particular query finds cases of a (Column1, Column2....ColumnN) with a (Column1) but it could be pretty easily modified to do a multiple left side key required. It's all filtering, just add [3] and the appropriate group by/max and you're set. And once again, yes, you could do this without a temp table, but the temp table's kind of interesting on its own.第二段怎么翻译，看不懂了，谢谢
At Kim Tripp's PASS session on indexing and covered indexes, she more-or-less proved that wider is better for indexes, within reason and searchability. It's better to have the index (Column1, Column2, Column3) than the index (Column1) and then a (Column1, Column2) and then a (Column1, Column2, Column3). And the first two of that set are redundant and don't provide any significant performance improvement over a wider index in most applications.Finding cases where you have multiple indexes like that isn't a trivial problem. Here's the answer I came up with. This particular query finds cases of a (Column1, Column2....ColumnN) with a (Column1) but it could be pretty easily modified to do a multiple left side key required. It's all filtering, just add [3] and the appropriate group by/max and you're set. And once again, yes, you could do this without a temp table, but the temp table's kind of interesting on its own.第二段怎么翻译，看不懂了，谢谢
我英语也不太好,希望理解的对...