Ok, 4 hours of coding and only 6 hours of searching... and I'm no better off than when I started. Here's my problem. I have a table (tmpShell) and it has 12 columns. It's a basic table with no constraints - used for temporary reporting. As we insert data, I have to extract an ID number (PatientId) and all column NAMES where the value for that PatientId is null.
Example:
PatientId Fname Lname DOB 123455 Sam NULL NULL
2345455 NULL Doe 1/1/1980
09172349 John Jone NULL
What I want to return is:
PatientId ErrorMsg 123455 Lname,DOB
2345455 Fname
09172349 DOB
Of course, if all columns have a value, the errormsg would be null.
I have tried and failed about 300 different pieces of code, but this appear to be the closest I can get. Unfortunately, this just returns EVERY column, not the nulls.
ALTER PROC [sp_aaShowAllNullColumns]
@tableName VARCHAR(255)
AS
BEGIN
SET NOCOUNT ON;
DECLARE @sql NVARCHAR(4000);
DECLARE @cols NVARCHAR(4000);
DECLARE @tcols TABLE ( [colbit] NVARCHAR(255) );
--DECLARE @tablename VARCHAR(255) = 'tmpShell';
INSERT @tcols
SELECT
'count(' + [columns].[name] + ') as ' + [columns].[name] + ', ' AS [colbit]
FROM
[sys].[columns]
WHERE
[columns].[object_id] = OBJECT_ID(@tableName);
SELECT
@cols = COALESCE(@cols, ', ', '') + [@tcols].[colbit]
FROM
@tcols;
SELECT
@cols = SUBSTRING(@cols, 1, ( LEN(@cols) - 1 ));
SELECT
@cols = ISNULL(@cols, '');
SELECT
@sql = 'select patientid, count(*) as Rows' + @cols + ' from ' + @tableName + ' group by patientid having count(*) > 0';
CREATE TABLE [tmpShell2]
(
[patientid] VARCHAR(15)
,[Rows] CHAR(2)
,[Rn] CHAR(2)
,[patId] CHAR(2)
,[fname] CHAR(2)
,[lname] CHAR(2)
,[dob] CHAR(2)
,[addr1] CHAR(2)
,[city] CHAR(2)
,[state] CHAR(2)
,[zip] CHAR(2)
,[country] CHAR(2)
,[psite] CHAR(2)
,[csite] CHAR(2)
,[ssite] CHAR(2)
,[scode] CHAR(2)
,[sfid] CHAR(2)
,[taskid] CHAR(2)
,[errormsg] CHAR(2)
);
INSERT INTO [tmpShell2]
EXEC [sys].[sp_executesql]
@sql;
DECLARE @tbl VARCHAR(255) = 'tmpShell2';
SELECT DISTINCT
[TS].[patientid]
, STUFF((
SELECT DISTINCT
', ' + [C].[name]
FROM
[tmpShell2] AS [TS2]
JOIN [sys].[columns] AS [C]
ON [C].[object_id] = OBJECT_ID(@tbl)
WHERE
[C].[name] NOT IN ( 'SFID', 'TaskId', 'ErrorMsg' )
AND [C].[name] IS NOT NULL
FOR
XML PATH('')
), 1, 1, '')
FROM
[tmpShell2] AS [TS];
DROP TABLE [dbo].[tmpShell2];
END;
GO
EXEC [sp_aaShowAllNullColumns]
'tmpShell';
</pre>
If you need to list all rows where all the column values are NULL , then i'd use the COLLATE function. This takes a list of values and returns the first non-null value. If you add all the column names to the list, then use IS NULL , you should get all the rows containing only nulls.
Use <=> (null-safe equality operator) negated comparison which returns FALSE in case one of the operands is null but TRUE when both are null and both operands have equal non-null values.
Null values do not behave like other values. For example, if you ask Db2 whether a null value is larger than a given known value, the answer is UNKNOWN. If you then ask Db2 whether a null value is smaller than the same known value, the answer is still UNKNOWN.
I think you over complicated things.
You can try using CASE EXPRESSION
:
SELECT t.patientID,
CASE WHEN t.fname is NULL THEN 'Fname,' ELSE '' END +
CASE WHEN t.Lname is NULL THEN 'Lname,' ELSE '' END +
CASE WHEN t.DOB is NULL THEN 'DOB,' ELSE '' END
..... as ErrorMsg
FROM YourTable t
This will result with an unnecessary comma in the end of the errorMsg
, to handle it you can do this:
REPLACE(CASE... +
CASE... +
CASE WHEN t.DOB is NULL THEN 'DOB,' ELSE '' END
..... + ' ') ', ','') as ErrorMsg
This will make the last comma unique because it will have a space concatenate to it and will make sure only it will be deleted .
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With