将大量（几乎100个）nvarchar列转换为浮点数的最快方法是什么？
目前我有一堆TRY_PARSE（列名AS FLOAT）语句，每列一个。
如果列不解析为float，我想要NULL。没有错误通知，没有额外的情报。
有更快的方法吗？如果是这样，会是什么？
SQL Server 2012
只是一个简短的测试，以显示TRY_PARSE和TRY_CAST之间的性能差异
我的结果
TRY_CAST 327 msTRY_PARSE 17784 ms这意味着，TRY_CAST比TRY_PARSE快约50倍（顺便说一句：TRY_CONVERT在表现上与TRY_CAST几乎相同）。
试试看并发布你的结果;-)
SET LANGUAGE ENGLISH;
DECLARE @tbl TABLE(SomeFloatAsString VARCHAR(100),fl2 VARCHAR(100),fl3 VARCHAR(100),fl4 VARCHAR(100),fl5 VARCHAR(100),fl6 VARCHAR(100),fl7 VARCHAR(100),fl8 VARCHAR(100),fl9 VARCHAR(100),fl10 VARCHAR(100));
DECLARE @i INT=0;
WHILE @i<100000
BEGIN
INSERT INTO @tbl VALUES(RAND(),RAND(),RAND(),RAND(),RAND(),RAND(),RAND(),RAND(),RAND(),RAND());
SET @i+=1;
END
DECLARE @d DATETIME2 = SYSUTCDATETIME();
SELECT TRY_CAST(SomeFloatAsString AS FLOAT) fl1
,TRY_CAST(fl2 AS FLOAT) fl2
,TRY_CAST(fl3 AS FLOAT) fl3
,TRY_CAST(fl4 AS FLOAT) fl4
,TRY_CAST(fl5 AS FLOAT) fl5
,TRY_CAST(fl6 AS FLOAT) fl6
,TRY_CAST(fl7 AS FLOAT) fl7
,TRY_CAST(fl8 AS FLOAT) fl8
,TRY_CAST(fl9 AS FLOAT) fl9
,TRY_CAST(fl10 AS FLOAT)fl10
INTO #tmp1
FROM @tbl
SELECT DATEDIFF(MILLISECOND,@d,SYSUTCDATETIME())
DECLARE @d2 DATETIME2 = SYSUTCDATETIME();
SELECT TRY_PARSE(SomeFloatAsString AS FLOAT) fl1
,TRY_PARSE(fl2 AS FLOAT) fl2
,TRY_PARSE(fl3 AS FLOAT) fl3
,TRY_PARSE(fl4 AS FLOAT) fl4
,TRY_PARSE(fl5 AS FLOAT) fl5
,TRY_PARSE(fl6 AS FLOAT) fl6
,TRY_PARSE(fl7 AS FLOAT) fl7
,TRY_PARSE(fl8 AS FLOAT) fl8
,TRY_PARSE(fl9 AS FLOAT) fl9
,TRY_PARSE(fl10 AS FLOAT)fl10
INTO #tmp2
FROM @tbl
SELECT DATEDIFF(MILLISECOND,@d2,SYSUTCDATETIME());
Shnugo发表评论说TRY_CAST应该更快，所以我认为有什么比测试更好的方式，而且从我做过的“快速”测试我同意。
首先，简短的设置来创建一些数据：
USE Sandbox;
GO
DECLARE @SQL nvarchar(MAX);
CREATE TABLE TallyTable (I int);
WITH N AS(
SELECT N
FROM (VALUES(NULL),(NULL),(NULL),(NULL),(NULL),(NULL),(NULL),(NULL),(NULL),(NULL)) N(N)),
Tally AS(
SELECT ROW_NUMBER() OVER (ORDER BY (SELECT NULL)) AS I
FROM N N1
CROSS JOIN N N2
CROSS JOIN N N3)
INSERT INTO dbo.TallyTable (I)
SELECT I
FROM Tally;
SELECT @SQL =
N'CREATE TABLE SampleTable (' +
STUFF((SELECT TOP 100
N',' + NCHAR(10) +
N' ' + QUOTENAME(CONCAT(N'Col',T.I)) + N' varchar(10)'
FROM TallyTable T
FOR XML PATH(N'')),1,1,N'') + NCHAR(10) + N');';
EXEC sp_executesql @SQL;
SELECT @SQL =
N'INSERT INTO SampleTable' + NCHAR(10) +
N'VALUES(' + STUFF((SELECT TOP 100 ',0'
FROM TallyTable T
FOR XML PATH(N'')),1,1,N'') + N');';
EXEC sp_executesql @SQL;
GO
这将创建一个包含100列的表，其值为“0”。现在，如果我们尝试TRY_CAST和TRY_PARSE：
SET STATISTICS TIME ON;
PRINT 'TRY_PARSE';
SELECT TRY_PARSE(Col1 AS float),
TRY_PARSE(Col2 AS float),
TRY_PARSE(Col3 AS float),
TRY_PARSE(Col4 AS float),
TRY_PARSE(Col5 AS float),
TRY_PARSE(Col6 AS float),
TRY_PARSE(Col7 AS float),
TRY_PARSE(Col8 AS float),
TRY_PARSE(Col9 AS float),
TRY_PARSE(Col10 AS float),
TRY_PARSE(Col11 AS float),
TRY_PARSE(Col12 AS float),
TRY_PARSE(Col13 AS float),
TRY_PARSE(Col14 AS float),
TRY_PARSE(Col15 AS float),
TRY_PARSE(Col16 AS float),
TRY_PARSE(Col17 AS float),
TRY_PARSE(Col18 AS float),
TRY_PARSE(Col19 AS float),
TRY_PARSE(Col20 AS float),
TRY_PARSE(Col21 AS float),
TRY_PARSE(Col22 AS float),
TRY_PARSE(Col23 AS float),
TRY_PARSE(Col24 AS float),
TRY_PARSE(Col25 AS float),
TRY_PARSE(Col26 AS float),
TRY_PARSE(Col27 AS float),
TRY_PARSE(Col28 AS float),
TRY_PARSE(Col29 AS float),
TRY_PARSE(Col30 AS float),
TRY_PARSE(Col31 AS float),
TRY_PARSE(Col32 AS float),
TRY_PARSE(Col33 AS float),
TRY_PARSE(Col34 AS float),
TRY_PARSE(Col35 AS float),
TRY_PARSE(Col36 AS float),
TRY_PARSE(Col37 AS float),
TRY_PARSE(Col38 AS float),
TRY_PARSE(Col39 AS float),
TRY_PARSE(Col40 AS float),
TRY_PARSE(Col41 AS float),
TRY_PARSE(Col42 AS float),
TRY_PARSE(Col43 AS float),
TRY_PARSE(Col44 AS float),
TRY_PARSE(Col45 AS float),
TRY_PARSE(Col46 AS float),
TRY_PARSE(Col47 AS float),
TRY_PARSE(Col48 AS float),
TRY_PARSE(Col49 AS float),
TRY_PARSE(Col50 AS float),
TRY_PARSE(Col51 AS float),
TRY_PARSE(Col52 AS float),
TRY_PARSE(Col53 AS float),
TRY_PARSE(Col54 AS float),
TRY_PARSE(Col55 AS float),
TRY_PARSE(Col56 AS float),
TRY_PARSE(Col57 AS float),
TRY_PARSE(Col58 AS float),
TRY_PARSE(Col59 AS float),
TRY_PARSE(Col60 AS float),
TRY_PARSE(Col61 AS float),
TRY_PARSE(Col62 AS float),
TRY_PARSE(Col63 AS float),
TRY_PARSE(Col64 AS float),
TRY_PARSE(Col65 AS float),
TRY_PARSE(Col66 AS float),
TRY_PARSE(Col67 AS float),
TRY_PARSE(Col68 AS float),
TRY_PARSE(Col69 AS float),
TRY_PARSE(Col70 AS float),
TRY_PARSE(Col71 AS float),
TRY_PARSE(Col72 AS float),
TRY_PARSE(Col73 AS float),
TRY_PARSE(Col74 AS float),
TRY_PARSE(Col75 AS float),
TRY_PARSE(Col76 AS float),
TRY_PARSE(Col77 AS float),
TRY_PARSE(Col78 AS float),
TRY_PARSE(Col79 AS float),
TRY_PARSE(Col80 AS float),
TRY_PARSE(Col81 AS float),
TRY_PARSE(Col82 AS float),
TRY_PARSE(Col83 AS float),
TRY_PARSE(Col84 AS float),
TRY_PARSE(Col85 AS float),
TRY_PARSE(Col86 AS float),
TRY_PARSE(Col87 AS float),
TRY_PARSE(Col88 AS float),
TRY_PARSE(Col89 AS float),
TRY_PARSE(Col90 AS float),
TRY_PARSE(Col91 AS float),
TRY_PARSE(Col92 AS float),
TRY_PARSE(Col93 AS float),
TRY_PARSE(Col94 AS float),
TRY_PARSE(Col95 AS float),
TRY_PARSE(Col96 AS float),
TRY_PARSE(Col97 AS float),
TRY_PARSE(Col98 AS float),
TRY_PARSE(Col99 AS float),
TRY_PARSE(Col100 AS float)
FROM dbo.SampleTable;
PRINT 'TRY_CAST';
SELECT TRY_CAST(Col1 AS float),
TRY_CAST(Col2 AS float),
TRY_CAST(Col3 AS float),
TRY_CAST(Col4 AS float),
TRY_CAST(Col5 AS float),
TRY_CAST(Col6 AS float),
TRY_CAST(Col7 AS float),
TRY_CAST(Col8 AS float),
TRY_CAST(Col9 AS float),
TRY_CAST(Col10 AS float),
TRY_CAST(Col11 AS float),
TRY_CAST(Col12 AS float),
TRY_CAST(Col13 AS float),
TRY_CAST(Col14 AS float),
TRY_CAST(Col15 AS float),
TRY_CAST(Col16 AS float),
TRY_CAST(Col17 AS float),
TRY_CAST(Col18 AS float),
TRY_CAST(Col19 AS float),
TRY_CAST(Col20 AS float),
TRY_CAST(Col21 AS float),
TRY_CAST(Col22 AS float),
TRY_CAST(Col23 AS float),
TRY_CAST(Col24 AS float),
TRY_CAST(Col25 AS float),
TRY_CAST(Col26 AS float),
TRY_CAST(Col27 AS float),
TRY_CAST(Col28 AS float),
TRY_CAST(Col29 AS float),
TRY_CAST(Col30 AS float),
TRY_CAST(Col31 AS float),
TRY_CAST(Col32 AS float),
TRY_CAST(Col33 AS float),
TRY_CAST(Col34 AS float),
TRY_CAST(Col35 AS float),
TRY_CAST(Col36 AS float),
TRY_CAST(Col37 AS float),
TRY_CAST(Col38 AS float),
TRY_CAST(Col39 AS float),
TRY_CAST(Col40 AS float),
TRY_CAST(Col41 AS float),
TRY_CAST(Col42 AS float),
TRY_CAST(Col43 AS float),
TRY_CAST(Col44 AS float),
TRY_CAST(Col45 AS float),
TRY_CAST(Col46 AS float),
TRY_CAST(Col47 AS float),
TRY_CAST(Col48 AS float),
TRY_CAST(Col49 AS float),
TRY_CAST(Col50 AS float),
TRY_CAST(Col51 AS float),
TRY_CAST(Col52 AS float),
TRY_CAST(Col53 AS float),
TRY_CAST(Col54 AS float),
TRY_CAST(Col55 AS float),
TRY_CAST(Col56 AS float),
TRY_CAST(Col57 AS float),
TRY_CAST(Col58 AS float),
TRY_CAST(Col59 AS float),
TRY_CAST(Col60 AS float),
TRY_CAST(Col61 AS float),
TRY_CAST(Col62 AS float),
TRY_CAST(Col63 AS float),
TRY_CAST(Col64 AS float),
TRY_CAST(Col65 AS float),
TRY_CAST(Col66 AS float),
TRY_CAST(Col67 AS float),
TRY_CAST(Col68 AS float),
TRY_CAST(Col69 AS float),
TRY_CAST(Col70 AS float),
TRY_CAST(Col71 AS float),
TRY_CAST(Col72 AS float),
TRY_CAST(Col73 AS float),
TRY_CAST(Col74 AS float),
TRY_CAST(Col75 AS float),
TRY_CAST(Col76 AS float),
TRY_CAST(Col77 AS float),
TRY_CAST(Col78 AS float),
TRY_CAST(Col79 AS float),
TRY_CAST(Col80 AS float),
TRY_CAST(Col81 AS float),
TRY_CAST(Col82 AS float),
TRY_CAST(Col83 AS float),
TRY_CAST(Col84 AS float),
TRY_CAST(Col85 AS float),
TRY_CAST(Col86 AS float),
TRY_CAST(Col87 AS float),
TRY_CAST(Col88 AS float),
TRY_CAST(Col89 AS float),
TRY_CAST(Col90 AS float),
TRY_CAST(Col91 AS float),
TRY_CAST(Col92 AS float),
TRY_CAST(Col93 AS float),
TRY_CAST(Col94 AS float),
TRY_CAST(Col95 AS float),
TRY_CAST(Col96 AS float),
TRY_CAST(Col97 AS float),
TRY_CAST(Col98 AS float),
TRY_CAST(Col99 AS float),
TRY_CAST(Col100 AS float)
FROM dbo.SampleTable;
SET STATISTICS TIME OFF;
以上结果是微不足道的，但它仍然表明：
TRY_PARSE
SQL Server Execution Times:
CPU time = 0 ms, elapsed time = 0 ms.
(1 row affected)
SQL Server Execution Times:
CPU time = 0 ms, elapsed time = 4 ms.
TRY_CAST
SQL Server Execution Times:
CPU time = 0 ms, elapsed time = 0 ms.
(1 row affected)
SQL Server Execution Times:
CPU time = 0 ms, elapsed time = 0 ms.
这是一个4毫秒不同的1列100列。那么让我们调高音量（这需要一段时间才能运行）：
DECLARE @SQL nvarchar(MAX);
SELECT @SQL =
N'INSERT INTO SampleTable' + NCHAR(10) +
N'VALUES' +STUFF((SELECT N',' + NCHAR(10) +
N' (' + STUFF((SELECT TOP 100 ',0'
FROM TallyTable T
FOR XML PATH(N'')),1,1,N'') + N')'
FROM TallyTable T
FOR XML PATH(N'')),1,1,N'') + N';';
SELECT @SQL;
EXEC sp_executesql @SQL;
EXEC sp_executesql @SQL;
EXEC sp_executesql @SQL;
EXEC sp_executesql @SQL;
EXEC sp_executesql @SQL;
--Now we have 5001 rows
现在我们有5,001行和100列。现在，如果我们再次运行上述...
TRY_PARSE
SQL Server Execution Times:
CPU time = 0 ms, elapsed time = 0 ms.
(5001 rows affected)
SQL Server Execution Times:
CPU time = 14672 ms, elapsed time = 18690 ms.
TRY_CAST
SQL Server Execution Times:
CPU time = 0 ms, elapsed time = 0 ms.
(5001 rows affected)
SQL Server Execution Times:
CPU time = 125 ms, elapsed time = 671 ms.
惊人的差异18019ms（即18秒，或快近28倍！）。
所以，答案是，使用TRY_CAST（或TRY_CONVERT，我相信它会一样快）。
--cleanup
DROP TABLE dbo.TallyTable;
DROP TABLE SampleTable;