Not getting results expected from a Group by query

  • Hi All,

    Hope you can shed some light. I'm a little puzzled right now. I have the following query:

    SELECT 'Independance' as SiteName, COUNT(att.AttendDate) AS Attendcount,

    s.ShelterName,

    (Select COUNT(PatientID) as PCount from PATIENTS where HomelessShelter = s.ShelterId and activeclient > 0) as PatientCount

    FROM PATIENTS p RIGHT OUTER JOIN

    APGAttendance Att ON p.PatientID = att.PatientID LEFT OUTER JOIN

    Shelters s ON p.HomelessShelter = s.ShelterId

    WHERE ((p.HomelessShelter > 0) )

    AND (att.AttendDate BETWEEN '10/29/2012' AND '11/27/2012')

    --and p.ActiveClient > 0

    GROUP BY s.ShelterName,s.ShelterId

    Some info on the Query. I need to know the total number of services that was provided to all clients in Shelters. And I also wanted to put the total number of clients that are actually in a particular shelter. Say I have 10 clients in shelter1 who have did 30 services within a date range given.

    The Problem. I happen to notice the issue because of the total number of client in a shelter count. In one of the totals it give say 58 total visits for say shelter12, but the Client count is 0. I checked the database tables to make sure that there was not any clients assigned to that shelter and there was no assingment to that shelter. So why do I still get a count for that shelter?

    Thanks for any input.

  • Michael, read this article, then post some sample data and ddl. http://www.sqlservercentral.com/articles/Best+Practices/61537/. It's hard to figure out what you are trying to get at without seeing what you are seeing. I would probably get rid of the right outer join if I were you. There is nothing wrong with it, but it is easy to become confused if you are having left and right joins in the same query. Once you post the data and ddl, I'm sure you will get a good solution.

    Greg
    _________________________________________________________________________________________________
    The glass is at one half capacity: nothing more, nothing less.

  • All I can see from what you've posted so far is the check for:

    activeclient > 0

    in the subquery (patient count) but not in the outer query (attendance count).

    It seems possible to me that the patients that made up the attendance count are no longer active in that shelter?!

    SQL DBA,SQL Server MVP(07, 08, 09) "It's a dog-eat-dog world, and I'm wearing Milk-Bone underwear." "Norm", on "Cheers". Also from "Cheers", from "Carla": "You need to know 3 things about Tortelli men: Tortelli men draw women like flies; Tortelli men treat women like flies; Tortelli men's brains are in their flies".

  • Here are the DDL for the tables

    ====Patients====

    [PatientID] [varchar](20) NOT NULL, --Primary Key

    [LastName] [nvarchar](25) NULL,

    [FirstName] [nvarchar](25) NULL,

    [AKA] [varchar](50) NULL,

    [ClinicianId] [int] NULL,

    [MEDICAID_ID] [nvarchar](8) NULL,

    [PendingMedicaid] [tinyint] NULL,

    [PrivatePatient] [tinyint] NULL,

    [MedicaidEffectiveDate] [datetime] NULL,

    [MedicaidTermDate] [datetime] NULL,

    [DOB] [datetime] NULL,

    [SEX] [char](1) NULL,

    [SSN] [varchar](12) NULL,

    [ActiveClient] [smallint] NULL,

    [HomelessShelter] smallint null

    Remaining fields left out

    ===APGAttendance====

    [PatientId] [varchar](20) NOT NULL, --Primary

    [AttendDate] [datetime] NOT NULL, --Primary

    [CTPCodes] [varchar](10) NOT NULL, --Primary

    [CheckinBySecurty] [bit] NULL,

    [SecurityTime] [varchar](12) NULL,

    [GroupTime] [varchar](12) NULL,

    [LocationID] [int] NULL,

    [ClinicianID] [int] NULL,

    Remaining fields left out to save space

    CREATE TABLE [dbo].[Shelters](

    [ShelterId] [int] IDENTITY(1,1) NOT NULL, --FOREIGN key to HomelessShelter

    [ShelterName] [varchar](50) NULL,

    [ShelterLocation] [varchar](100) NULL,

    [SortOrder] [tinyint] NULL,

    [isFreedomRes] [bit] NULL,

    CONSTRAINT [PK_Shelters] PRIMARY KEY CLUSTERED

    (

    [ShelterId] ASC

    )WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]

    ) ON [PRIMARY]

    Here is the data for Shelter table:

    ShelterId Sheltername

    ----------- --------------------------------------------------

    1 Freedom Residence I

    2 Freedom Residence II

    3 Freedom Residence III

    4 Freedom Residence IV

    5 Freedom Residence V

    6 Freedom Residence VI

    7 Freedom Residence VII

    8 New Sight to Life

    9 Jastian Management

    10 Freedom Residence VIII

    11 Freedom Residence VIIII

    12 Joining Hands

    13 Jireh Place

    14 #1 Marketing/RYB/Top of the Hob

    15 Uplifting Men

    16 No Matter What/Rockaway

    17 Forever USA

    18 Well Being/WINS Enterprise

    19 Steps to Success

    20 Freedom Residence X

    21 Freedom Residence XI

    22 Freedom Residence XII

    23 Freedom Residence XVI

    24 Liberation

    25 Freedom Residence XVII

    26 Freedom Residence XVIII

    27 Freedom Residence XVIIII

    28 Freedom Residence XX

    29 Freedom Residence XXI

    30 Freedom Residence XXII

    31 Freedom Residence XIII

    32 Freedom Residence XIIII

    33 Freedom Residence XV

    The Results from the Query in question.

    NarcoSite Attendcount ShelterName PatientCount

    ------------ ----------- --------------------------- ------------

    Independance 120 Freedom Residence I 10

    Independance 212 Freedom Residence II 27

    Independance 250 Freedom Residence III 37

    Independance 96 Freedom Residence IV 16

    Independance 165 Freedom Residence V 21

    Independance 170 Freedom Residence VI 19

    Independance 320 Freedom Residence VIII 41

    Independance 156 Freedom Residence XI 13

    Independance 193 Freedom Residence XII 24

    Independance 319 Freedom Residence XVI 38

    Independance 58 Freedom Residence XVIII 0 <--Row in question

    Independance 352 Freedom Residence XVIIII 46

    Independance 14 Freedom Residence XIIII 1

    BWT, I'm currently have the field Sitename in the query to distinguish which site the data is coming from and will be used in a report the data is feeding. This is part of a Union all query that retrieves the data from several servers.

    In reference to the row in question, I did check the patient table and know that there are no records that have the HomelessShelter field set to 26. Not sure why it would display 58 as the count.

    I did try and change around the join clause with a left join instead of an outer join. Pretty much the same results though.

    I understand that the subquery may not be the best way to get the Patient Count, but was not sure of a better way to get this info without using a function, which would have to use the same type query to get the count anyways, so for now I had it within the parent query. If there is a better way, I will be happy to do it.

    Thanks for the suggestions so far.

  • Here are the DDL for the tables

    Here is the data for Shelter table:

    Michael, did you read the link I posted yesterday? *All of it*? You are heading in the right direction with respect to the ddl and sample data, but you still need to do a little more. We should be able to copy and paste your ddl and sample data and run it as is, without having to type any create and insert statements. Once you do this, you will be more likely to get help. Now, there are folks who will do this for you (sometimes I do, but don't have time today), so you could wait and see if someone does, or you could help us help you.

    Greg
    _________________________________________________________________________________________________
    The glass is at one half capacity: nothing more, nothing less.

Viewing 5 posts - 1 through 4 (of 4 total)

You must be logged in to reply to this topic. Login to reply