Here is the scenario. We have a column of type IMAGE in a table that holds compressed data. This column can be null. In a different version, we moved this data to a separate table, still an image column.
We have a stored procedure that is checking where the data is, in the original table or the new table. To do that, it selects the value into a VARBINARY(MAX) variable, checks if it is null, if so grabs it from the new table.
Something strange is going on with the checks. For a few rows in the database, the column displays as NULL, but when cast to VARBINARY(MAX) it ends up being an empty value of 0x. This is only for 4 rows out of 426, all others evaluate to NULL fine. If you check for null in the table itself, like WHERE <column> IS NULL it evalulates to true, but if you do WHERE CAST(<column> as VARBINARY(MAX)) IS NULL it evaluates to false.
I can change around my stored procedure so that it handles correctly, but I'm a little baffled as to what is going on here. Any ideas?
This is on SQL Server 2005.
***edit to try and include a clearer example
e.g.
SELECT a
FROM Table
WHERE a IS NULL
SELECT a
FROM Table
WHERE CAST(a as VARBINARY(MAX)) IS NULL
the first query returns the result with NULL, the second query doesn't return anything.
Similarily, the stored procedure looks something like this:
DECLARE #CompressedData VARBINARY(MAX)
SELECT #CompressedData = imageColumn
FROM originalTable
WHERE ID = #ID
IF (#CompressedData IS NULL)
BEGIN
SELECT #CompressedData = imageColumn
FROM newTable
WHERE ID = #ID
END
In that case, #CompressedData does not evaluate to NULL since it is getting a value of 0x
Related
I have a stored procedure that includes a table variable which updates from other table variables. ISNULL() is not working in the below code. It returns the proper value if there is something in TR.TotalRequestCnt, but NULL if there is no value in the TotalRequest table for this column.
UPDATE #OutputTable
SET TotalRequestCount = ISNULL(TR.TotalRequestCnt, 0)
FROM #TotalRequest TR
JOIN #OutputTable OT
ON TR.Document = OT.Document
Values of Document are identifying INTs (5577,5575, 5574).
#TotalRequest values are all INT:
5577 NULL NULL
5575 NULL NULL
5574 2 1
I have also tried using COALESCE() instead of ISNULL() to no success.
I found the answer.
No update is being performed in the row at all because there are no values so ISNULL never has the chance to fire.
I'm using C# to write to a SQL Compact Edition 3.5 database. I got a table containing e-mail addresses and names for each address.
MailRecipientAddressID int primary key identity(1,1) not null,
Address nvarchar(4000),
Name nvarchar(4000)
In this table I want every address-name combination to be unique. In some cases it's possible that either Address or Name is NULL in a row. Before inserting new rows into this table, I'm using a SELECT query to check if there is an existing row matching the one I want to insert. When using this query
SELECT MailRecipientAddressID FROM MailRecipientAddress WHERE Address = #Address AND Name = #Name
I will not find existing rows with NULL values in one column (see here).
Now I got this query, which works and kind of solves my problem
SELECT MailRecipientAddressID FROM MailRecipientAddress WHERE ISNULL(Address, '') = ISNULL(#Address, '') AND ISNULL(Name, '') = ISNULL(#Name, '')
but even though it is no problem in my case that NULL and empty string values are handled equally, I do not like this solution. I think it's kind of hackish. Is there a better approach
to apply a filter on a SELECT statement with parameters which can contain NULL
which works on SQL CE
Edit
I do not understand why, but my query works with SQL Management Studio but it does not in my application (see here). To correct my own approach I would need to use COALESCE.
I don't like the option to replace my NULL values with empty strings because I think it would be kind of inconsequent to set a value at a place where I got no value or is my understanding of this design question wrong?
The best solution is a constraint on the table that prevents duplicates from going into the table. You can put one in with a unique index:
create unique index idx_MailRecipientAddress_address_name on MailRecipientAddress(Address, Name);
This will generate an error on the insert, which you would then need to catch.
However, this is only a partial solution, because NULL values do not count as duplicates. You might solve your overall problem by not allowing NULL values in the field at all. Instead, represent no data using empty strings. Note: I wouldn't normally recommend this. In SQL, NULL means "unknown" and by the definition of the language, two "unknown" values are not equal. However, you seem to want them to be equal.
As for SQL, yours is okay, but it equates NULL and the empty string. An explicit check is more accurate:
WHERE (Address = #Address or Address is null and #Address is null) and
(Name = #Name or Name is null and #Name is null)
#George
if Parameter value is Null and column value is not null then "(Address = #Address or Address is NULL) returns false "
if Parameter value is Null and column value is null then "(Address = #Address or Address is NULL) returns true"
if Parameter value is Not Null and column value is null then "(Address = #Address or Address is NULL) returns true"
if Parameter value is Not Null and column value is Not null and if matches then "(Address = #Address or Address is NULL) returns true otherwise false"
SELECT MailRecipientAddressID FROM MailRecipientAddress WHERE (Address = #Address or Address is NULL) AND (Name = #Name or Name is NULL)
How can I write a SQL stored procedure where I want the parameters to be optional in the select statement?
try this.. Make the SPs input parameters that control the filtering optional, witrh default values of null. In each select statement's Where clause, write the predicate like this:
Create procedure MyProcedure
#columnNameValue [datatype] = null
As
Select [stuff....]
From table
Where ColumnName = Coalesce(#columnNameValue , ColumnName)
this way if you do not include the parameter, or if you pass a null value for the parameter, the select statement will filter on where the column value is equal to itself, (effectively doing no filtering at all on that column.)
The only negative to this is that it prevents you from being able to pass a null as a meaningfull value to explicitly filter on only the nulls.... (i.e., Select only the rows where the value is null) Once the above technique has been adopted, you would need to add another parameter to implement that type of requirement. ( say, #GetOnlyNulls TinyInt = 0, or something similar)
Create procedure MyProcedure
#columnNameValue [datatype] = null,
#GetOnlyNulls Tinyint = 0
As
Select [stuff....]
From table
Where (ColumnName Is Null And #GetOnlyNulls = 1)
Or ColumnName = Coalesce(#columnNameValue , ColumnName)
I have two views with identical columns. One of the columns on the first view is generated 'on the fly' and set as NULL and the same column on the other view has values stored as varchar. I have a stored proc that looks like this:
ALTER PROCEDURE [dbo].[mi_GetLearners]
(#centrename nvarchar(200))
AS
SELECT [centrename]
,[Name]
,[Username] --generated on the fly as NULL
FROM [DB1].[dbo].[vw_Learners]
WHERE [centrename] = #centrename
UNION
SELECT [centrename]
,[Name]
,[Username] --values stored as varchar
FROM [Linked_Server].[DB2].[dbo].[vw_Learners]
WHERE [centrename] = #centrename
DB1 is on SQL Server 2008 R2
DB2 is on SQL server 2005
When I run the stored proc I get the following error:
Msg 245, Level 16, State 1, Line 1 Conversion failed when converting
the varchar value 'someusername' to data type int.
Why is it trying to convert the value to int datatype as the other column is set as NULL? If I instead change the second column from NULL to ' ' the stored proc works fine... I'm really baffled as to why a union between a varchar column and NULL column generated in a select statement would throw such an error... any ideas?
EDIT: I'm looking for an explanation rather than a solution...
EDIT 2: Running the following code:
CREATE VIEW vw_myview
AS
SELECT NULL AS MyColumn
EXECUTE sp_help vw_myview
Returns:
Type Column_name
int MyColumn
The problem is that NULL is (to within some wiggle room) a member of every data type.
When any SELECT query is being run, the contents of each column must be of one type, and only one type. When there are a mixture of values in a column (including NULLs), the type can obviously be determined by examining the types of the non-NULL values, and appropriate conversions are performed, as necessary.
But, when all rows contain NULL for a particular column, and the NULL hasn't been cast to a particular type, then there's no type information to use. So, SQL Server, somewhat arbitrarily, decides that the type of this column is int.
create view V1
as
select 'abc' as Col1,null as Col2
go
create view V2
as
select 'abc' as Col1,CAST(null as varchar(100)) as Col2
V1 has columns of type varchar(3) and int.
V2 has columns of type varchar(3) and varchar(100).
I'd expect the type of the field to be determined by the first SELECT in the union. You may try changing the order of your two selects or change to CAST(NULL AS VARCHAR(...)).
I think you need to cast to the same datatype:
---
AS
SELECT [centrename]
,[Name]
,CAST(NULL AS VARCHAR(MAX)) AS [Username]
FROM [DB1].[dbo].[vw_Learners]
WHERE [centrename] = #centrename
UNION
---
Please suppose that in SQL Server 2005, if you launch the following query:
SELECT CHICKEN_CODE FROM ALL_CHICKENS WHERE MY_PARAMETER = 'N123123123';
you obtain:
31
as result.
Now, I would like to write a function that, given a value for MY_PARAMETER, yields the corresponding value of CHICKEN_CODE, found in the table ALL_CHICKENS.
I have written the following stored function in SQL Server 2005:
ALTER FUNCTION [dbo].[determines_chicken_code]
(
#input_parameter VARCHAR
)
RETURNS varchar
AS
BEGIN
DECLARE #myresult varchar
SELECT #myresult = CHICKEN_CODE
FROM dbo.ALL_CHICKENS
WHERE MY_PARAMETER = #input_parameter
RETURN #myresult
END
But if I launch the following query:
SELECT DBO.determines_chicken_code('N123123123')
it yields:
NULL
Why?
Thank you in advance for your kind cooperation.
define the length of your varchar variables like this
varchar(100)
Without the 100 (or whatever length you choose) its lengh is 1 and the where clause will filter out the correct results.
Specify a length for your varchar (ex.: varchar(100)). Without length, varchar = 1 char.
As per other PS, You can store only one char in the #myresult because you have not specified any length, bcoz 1 char length is default for Varchar datatype.
Why we are getting NUll, not the first char:
If there are multiple records are filtered on basis of Where clause in ALL_CHICKENS table then the value of CHICKEN_CODE column is picked up from last row in ALL_CHICKENS table.
It seems that the last row has null value in CHICKEN_CODE column.
Specify a length for #input_parameter, #myresult as by default varchar lengh is 1.