SQL AVG() 为 3 列返回错误的结果 [英] SQL AVG() Returning the wrong result for 3 columns

问题描述

我正在编写一个查询，它应该给我一个 count() 和三个 avg().count() 工作得很好，但 avg() 函数返回错误的结果.我正在处理的数据如下所示:

I am writing a query that is supposed to give me a count() and three avg()'s. The count() works just fine, but the avg() functions are returning erroneous results. The data I am working with looks like this:

MD Name |    PT | Med Staff | LOS | DRG Bench | LOS - Bench
MCP     | 12345 | Ortho SX  |  5  |    4      |       1
MCP     | 25879 | Ortho SX  |  3  |    5      |      -2
MCP     | 98556 | Ortho SX  |  4  |    5      |      -1
... 

我想要的输出是:

MD Name | # PT | Med Staff | Avg LOS | Avg DRG Bench | AVG LOS - Bench
MCP     |   3  | Ortho SX  |    4    |       4.66    |       0

我得到的平均值结果不正确.我有一种情况，特别是我有以下情况:

I am getting incorrect results for the averages. I have one case in particular where I have the following:

MD Name | LOS    | Bench  | LOS - Bench
MCP     | 2.0000 | 1.8000 |    0.2000
MCP     | 1.0000 | 1.7000 |   -0.7000
MCP     | 25.0000| 4.9000 |   20.1000
MCP     | 4.0000 | 2.2000 |    1.8000

对于和 AVG LOS，我得到 9.000000 而不是 8.000000，对于 AVG Bench，我得到 2.780000 而不是 2.65，对于 LOS-Bench，我得到 6.220000 而不是 5.35，这些是显着差异，我必须是正确的两位小数.

For and AVG LOS I am getting 9.000000 instead of 8.000000, for the AVG Bench I am getting 2.780000 instead of 2.65 and for the LOS-Bench I am getting 6.220000 instead of 5.35, these are significant differences and I must be correct to two decimal places.

这是我使用的 SQL，SQL Server 2008

Here is the SQL I am using, SQL Server 2008

DECLARE @STARTDATE DATETIME
DECLARE @ENDATE DATETIME

SET @STARTDATE = '2013-05-01'
SET @ENDATE = '2013-05-31'

SELECT DISTINCT pv.pract_rpt_name AS 'PHYSICIAN'
, COUNT(DISTINCT vr.pt_id) AS '# PTS' 
--, pv.spclty_desc AS 'SPECIALTY'
, pv.med_staff_dept AS 'MED STAFF'
, AVG(vr.len_of_stay) AS 'LOS'
, AVG(vr.drg_std_days_stay) AS 'DRG LOS BENCH'
, AVG(vr.len_of_stay - vr.drg_std_days_stay) AS 'LOS - DRG BENCH'

FROM smsmir.vst_rpt vr
LEFT OUTER JOIN smsmir.pyr_plan pp <-- removed and fixed
ON vr.pt_id = pp.pt_id <-- removed and fixed
JOIN smsdss.pract_dim_v pv
ON vr.adm_pract_no = pv.src_pract_no

WHERE vr.adm_dtime BETWEEN @STARTDATE AND @ENDATE
AND vr.vst_type_cd = 'I'
AND pv.spclty_desc != 'NO DESCRIPTION'
--AND pv.spclty_desc NOT LIKE 'HOSPITALIST%'
AND vr.drg_std_days_stay IS NOT NULL
AND pv.pract_rpt_name != '?'
AND pv.orgz_cd = 's0x0'
AND pv.med_staff_dept IN (
'INTERNAL MEDICINE',
'FAMILY PRACTICE',
'SURGERY'
)
GROUP BY pv.pract_rpt_name, pv.med_staff_dept
ORDER BY pv.med_staff_dept, AVG(vr.len_of_stay - vr.drg_std_days_stay)DESC

感谢您的时间和努力.

推荐答案

您在查询中的 OUTER JOIN 可能会影响 AVG 函数操作的行数.如果您不需要它(而且我在您的查询中的其他地方都看不到该表被引用的任何地方)，请尝试将其删除.

The OUTER JOIN you have in the query may be affecting the number of rows that the AVG function is operating over. If you don't need it (and I can't see anywhere where that table is referenced elsewhere in your query) try removing it.

这篇关于SQL AVG() 为 3 列返回错误的结果的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！