What am I doing wrong with this subquery?

  • I am trying to return all values from the lookup table to columns on the same row for the Diagnosis code.

     

    Here is my query:

    SELECT distinct

    a.accession_no as "Accession Number",

    (ISNULL((select m.code from medical_code m where ai.icd9_id = m.id and ai.sort_ord = 1), '')) + '~' +

    (ISNULL((select m.code from medical_code m where ai.icd9_id = m.id and ai.sort_ord = 2), '')) + '~' +

    (ISNULL((select m.code from medical_code m where ai.icd9_id = m.id and ai.sort_ord = 3), '')) + '~' +

    (ISNULL((select m.code from medical_code m where ai.icd9_id = m.id and ai.sort_ord = 4), '')) + '~' +

    (ISNULL((select m.code from medical_code m where ai.icd9_id = m.id and ai.sort_ord = 5), '')) + '~' +

    (ISNULL((select m.code from medical_code m where ai.icd9_id = m.id and ai.sort_ord = 6), '')) + '~' +

    (ISNULL((select m.code from medical_code m where ai.icd9_id = m.id and ai.sort_ord = 7), '')) + '~' as "Diagnosis Code"

    FROM accession_2 a (NOLOCK)

    LEFT OUTER JOIN acc_icd9 ai on a.id = ai.acc_id

    WHERE

    a.accession_no in ('S22-01587')

     

    Here is my query, but it's returning 2 rows instead of 1 row with diagnosis code as code~code~code~code~code~code~code~

    or in this case code~code~~~~~

     

    Here are my results:

    Accession Number Diagnosis Code

    S22-01587 ~N17.9~~~~~~

    S22-01587 E11.21~~~~~~~

     

    Thank you for your assistance!

  • You provided no sample data, so no way to test.  But something like this should help.  If table "accession_2" does not contain duplicate "accession_number"s, you should be able to remove the "distinct".

    SELECT distinct

    a.accession_no as "Accession Number",

    diag_codes."Diagnosis Code"

    FROM accession_2 a WITH (NOLOCK)

    OUTER APPLY (
    SELECT STUFF(
    ISNULL(MAX(CASE WHEN ai.sort_ord = 1 THEN '~' + m.code END), '') +
    ISNULL(MAX(CASE WHEN ai.sort_ord = 2 THEN '~' + m.code END), '') +
    ISNULL(MAX(CASE WHEN ai.sort_ord = 3 THEN '~' + m.code END), '') +
    ISNULL(MAX(CASE WHEN ai.sort_ord = 4 THEN '~' + m.code END), '') +
    ISNULL(MAX(CASE WHEN ai.sort_ord = 5 THEN '~' + m.code END), '') +
    ISNULL(MAX(CASE WHEN ai.sort_ord = 6 THEN '~' + m.code END), '') +
    ISNULL(MAX(CASE WHEN ai.sort_ord = 7 THEN '~' + m.code END), '')
    , 1, 1, '') AS "Diagnosis Code"
    FROM medical_code m
    LEFT OUTER JOIN acc_icd9 ai on a.id = ai.acc_id AND m.id = ai.icd9_id AND ai.sort_ord BETWEEN 1 AND 7
    ) AS diag_codes("Diagnosis Code")

    WHERE

    a.accession_no in ('S22-01587')

    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".

  • >> I am trying to return all values from the lookup table to columns on the same row for the Diagnosis code. <<

    Please read any book, and I do mean any book, on relational databases. This is a complete violation of first normal form. It also violates the concept of a tiered architecture. The database tier in a client/server architecture is supposed to return the data in a standardized format the table. If you want to print it in colors or use a fancy font or concatenated together or whatever else, then you do that in a presentation layer.

    Again, you failed to post any DDL so we have no idea just what a mess this thing is. You might also start using the ANSI/ISO standard coalesce () instead of the old Sybase ISNULL ().

    Please post DDL and follow ANSI/ISO standards when asking for help. 

  • You might also want to investigate the aggregate function STRING_AGG(), which should make what I think you want very easy.

    https://docs.microsoft.com/en-us/sql/t-sql/functions/string-agg-transact-sql?view=sql-server-ver15

     

  • jcelko212 32090 wrote:

    This is a complete violation of first normal form. It also violates the concept of a tiered architecture.

    While that is true, this is a common requirement for medical data. You do not argue with the government if you want to be a Medicare or Medicaid provider.

    My question - ICD9?

     

  • It would obviously be easier to offer help if you'd only provide some kind of test data for us to play around with, and the desired outcome for the test data.

     

  • International classification of diseases amended – version 9

    This is the standard encoding used in medicine. However, I think they might be on version 11 by now. Some of the newer codes included "sucked through a jet engine.", "Bitten by a shark." and some bizarre new diseases.

    Please post DDL and follow ANSI/ISO standards when asking for help. 

  • Okay, based on your query and your query output I have tried to guess some of your data and created some test tables (I have used declared table variables):

    declare @accession_2 table (
    Id int,
    accession_no varchar(20)
    )
    insert into @accession_2 (Id, accession_no)
    values
    (1,'S22-01587'),
    (2,'S22-01587')

    declare @acc_icd9 table (
    acc_id int,
    icd9_id int,
    sort_ord int
    )

    insert into @acc_icd9 (acc_id, icd9_id, sort_ord)
    values
    (1,1,2),
    (2,2,1)

    declare @medical_code table (
    Id int,
    Code varchar(20)
    )

    insert into @medical_code (Id, Code)
    values
    (1,'E11.21'),
    (2,'N17.9')

    I obviously don't know if this is a correct representation of your data, but at least I can run your query and arrive at the same output as you:

    Accession Number     Diagnosis Code
    -------------------- ---------------------------------------------------------------------------------------------------------------------------------------------------
    S22-01587 ~E11.21~~~~~~
    S22-01587 N17.9~~~~~~~

    (2 rows affected)

    This lets me have some confidence in the data, and based on those data I have come up with this query:

    declare @AllowedSortId table (
    sortId tinyint
    )

    insert into @AllowedSortId (sortId)
    values
    (1),
    (2),
    (3),
    (4),
    (5),
    (6),
    (7)

    With Consolidated_accession AS (
    SELECT
    a.accession_no,
    asi.sortId,
    isnull(max(m.code),'') as Code

    FROM @accession_2 a
    cross join @AllowedSortId asi
    LEFT OUTER JOIN @acc_icd9 ai on a.id = ai.acc_id and ai.sort_ord=asi.sortId
    left outer join @medical_code m on m.Id=ai.icd9_id and ai.sort_ord = asi.sortId
    GROUP BY a.accession_no, asi.sortId
    )
    select
    accession_no as "Accession Number",
    string_agg(code,'~') WITHIN GROUP (ORDER BY sortId ASC) as "Diagnosis Code"
    from Consolidated_accession
    WHERE
    accession_no in ('S22-01587')
    group by accession_no
    order by accession_no

    The output:

    Accession Number     Diagnosis Code
    -------------------- ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
    S22-01587 N17.9~E11.21~~~~~

    Maybe this is closer to what you expect?

    As you can see I've added an additional table with the allowable sort_ids.

    • This reply was modified 2 years, 9 months ago by  kaj. Reason: Added column to join criteria for table medicalcode, and added comment about the additional table used
  • This was removed by the editor as SPAM

Viewing 9 posts - 1 through 8 (of 8 total)

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