Friday, October 26, 2007

Do you want to laugh or cry?

I don't know if I should laugh or cry after looking at this code. This could not have been written by a human right?

Anyway here it is, consider it a teaser. If you want to torture yourself and look at the whole query then go here: http://www.tek-tips.com/viewthread.cfm?qid=1421304&page=1



HAVING (((1 = 1)
AND (PP.Created >= ISNULL(NULL,'1/1/1900')
AND PP.Created < 1 =" 2)">= ISNULL(NULL,'1/1/1900')
AND (MIN(PV.Visit)) < DATEADD(d, 1, ISNULL(NULL,'1/1/3000')))
))



WHERE
(
(NULL IS NOT NULL
AND pv.DoctorId IN (NULL))
OR(NULL IS NULL)
)



CASE
WHEN '0' = 1 THEN df.ListName
WHEN '0' = 2 THEN df2.ListName
WHEN '0' = 3 THEN ic.ListName
ELSE NULL
END


Someone will red flag the post, here is the whole query.


/* New Patient Analysis */

SET NOCOUNT ON

CREATE TABLE #PatientTemp
(
[PatientProfileId] [int] NOT NULL ,
[Prefix] [varchar] (10) NULL ,
[First] [varchar] (30) NULL ,
[Middle] [varchar] (30) NULL ,
[Last] [varchar] (30) NULL ,
[Suffix] [varchar] (20) NULL ,
[Created] [datetime] NOT NULL
)

Insert #PatientTemp

SELECT PP.PatientProfileId, PP.Prefix, PP.First, PP.Middle, PP.Last, PP.Suffix,
CASE
WHEN 1 = 1 THEN PP.Created
WHEN 1 = 2 THEN Min(PV.Visit)
ELSE NULL
END As Created

FROM PatientVisit PV
INNER JOIN PatientProfile PP ON PP.PatientProfileId = PV.PatientProfileId

GROUP BY PP.PatientProfileId, PP.Prefix, PP.First, PP.Middle, PP.Last, PP.Suffix, PP.Created

HAVING (((1 = 1)AND
(PP.Created >= ISNULL(NULL,'1/1/1900') AND PP.Created < DATEADD(d, 1, ISNULL(NULL,'1/1/3000')))
))
OR
((1 = 2)AND
((MIN(PV.Visit) >= ISNULL(NULL,'1/1/1900') AND (MIN(PV.Visit)) < DATEADD(d, 1, ISNULL(NULL,'1/1/3000')))
))

Order By PP.First, PP.Last

SELECT dbo.FormatName(#PatientTemp.Prefix, #PatientTemp.First, #PatientTemp.Middle, #PatientTemp.Last, #PatientTemp.Suffix) AS Name,
CASE
WHEN 1 = 1 THEN #PatientTemp.Created
WHEN 1 = 2 THEN Min(pv.Visit)
ELSE NULL
END As Created,
COUNT(*) AS [Number Of Visits],
SUM(pva.OrigInsAllocation + pva.OrigPatAllocation) AS Charges,
SUM(pva.InsPayment + pva.PatPayment) AS Payments,
CASE
WHEN '0' = 1 THEN df.ListName
WHEN '0' = 2 THEN df2.ListName
WHEN '0' = 3 THEN ic.ListName
ELSE NULLA
END As Grouping

FROM PatientVisit pv
INNER JOIN PatientVisitAgg pva ON pv.PatientVisitId = pva.PatientVisitId
INNER JOIN #PatientTemp ON pv.PatientProfileId = #PatientTemp.PatientProfileId
INNER JOIN DoctorFacility df ON pv.DoctorId = df.DoctorFacilityId
LEFT JOIN DoctorFacility df1 ON pv.ReferringDoctorId = df1.DoctorFacilityId
LEFT JOIN DoctorFacility df2 ON pv.FacilityId = df2.DoctorFacilityId
LEFT JOIN InsuranceCarriers ic ON ic.InsuranceCarriersId = pv.PrimaryInsuranceCarriersId

WHERE
(
(NULL IS NOT NULL AND pv.DoctorId IN (NULL)) OR
(NULL IS NULL)
)
AND
(
(NULL IS NOT NULL AND pv.ReferringDoctorId IN (NULL)) OR
(NULL IS NULL)
)
AND
(
(NULL IS NOT NULL AND pv.FacilityId IN (NULL)) OR
(NULL IS NULL)
)
AND --Filter on insurance carrier
(
(NULL IS NOT NULL AND ic.InsuranceCarriersId IN (NULL)) OR
(NULL IS NULL)
)

GROUP BY #PatientTemp.Created,dbo.FormatName(#PatientTemp.Prefix, #PatientTemp .First, #PatientTemp.Middle, #PatientTemp.Last, #PatientTemp.Suffix),
pva.PatientProfileId,
CASE
WHEN '0' = 1 THEN df.ListName
WHEN '0' = 2 THEN df2.ListName
WHEN '0' = 3 THEN ic.ListName
ELSE NULL
END

HAVING (((1 = 1)AND
(#PatientTemp.Created >= ISNULL(NULL,'1/1/1900') AND #PatientTemp.Created < DATEADD(d, 1, ISNULL(NULL,'1/1/3000')))
))
OR
((1 = 2)AND
((MIN(pv.Visit) >= ISNULL(NULL,'1/1/1900') AND (MIN(pv.Visit)) < DATEADD(d, 1, ISNULL(NULL,'1/1/3000')))
))

ORDER BY dbo.FormatName(#PatientTemp.Prefix, #PatientTemp.First, #PatientTemp.Middle, #PatientTemp.Last, #PatientTemp .Suffix), #PatientTemp.Created
Drop table #PatientTemp

2 comments:

Matija Lah said...

This is either a joke or beyond WTF...

Since the original article is no longer available - where, how and why have you found this?

SQL said...

Matija,

It was not a joke, I knew it would be red-flagged that is why I grabbed it before that happened. It was on tek-tips
It looks like it is code generated from a tool where they always use WHERE 1=1 so that they always generate the AND part of the WHERE clause