As part of the company I am working for at the moment I need to create some database upgrade scripts to replace some work of a previous contractor.
The code before the following block runs, creates the new ID column, and then this script looks to populate the values and then drop some columns.
IF EXISTS (
SELECT *
FROM sys.columns
WHERE object_id = OBJECT_ID(N'[Central].[Core.Report].[ReportLessonComp]')
AND name = 'Name')
and
EXISTS (
SELECT *
FROM sys.columns
WHERE object_id = OBJECT_ID(N'[Central].[Core.Report].[ReportLessonComp]')
AND name = 'Code')
BEGIN
UPDATE
[Central].[Core.Report].[ReportLessonComp]
SET
CompetencyId = rc.Id
FROM
[Central].[Core.Report].[ReportLessonComp] rlc
INNER JOIN
[Core.Lookup].ReportCompetency rc
ON
rc.Code = rlc.Code and rc.Name = rlc.Name
ALTER TABLE [Central].[Core.Report].[ReportLessonComp] DROP COLUMN CODE
ALTER TABLE [Central].[Core.Report].[ReportLessonComp] DROP COLUMN [Name]
ALTER TABLE [Central].[Core.Report].[ReportLessonComp] DROP COLUMN [Description]
END
GO
When running the if exists \ not exists checks and then select getdate() this works perfeclty fine and gives me the result I expect.
However, when I run the code block above I get error
Msg 207, Level 16, State 1, Line 23
Invalid column name 'Code'.
Msg 207, Level 16, State 1, Line 23
Invalid column name 'Name'.
This script it part of a larger upgrade script and is used in a system calle RoundHouse https://github.com/chucknorris/roundhouse which is the system chosen by the company.
Prior to the above if exists check,
IF (SELECT COUNT(1) FROM sys.columns
WHERE OBJECT_ID = OBJECT_ID('[Central].[Core.Report].[ReportLessonComp]')
AND Name in ('Name','Code')) = 2
which also gave the same issue. I have five tables that I need to update and this is going to stop the team from working if I cant resolve this at my next PR
What can I do in order to stop this from causing the upgrade scripts to fail?
EDIT -- The reason I am linking on varchar fields also is because the previous developer did not create relationships between tables, and was just inserting strings into tables rather than relating by ID causing the potential for unlinked \ inconsistent data.
The table edit prior to this creates the new id column, and this script is getting the value and dropping columns that are no longer needed
SQL Server will parse the whole of the statement prior to execution, so the exists check does not protect you from the update being parsed. If the column has already been dropped, that makes the statement invalid and you get a parse error. The update statement would have to be executed as dynamic SQL, sp_execute basically so that the varchar of the update is not directly parsed.
For SQL Server 2016 and above the drop column can be protected a bit more as well:
ALTER TABLE [Central].[Core.Report].[ReportLessonComp] DROP COLUMN IF EXISTS CODE
Related
I want to write sql script that should check if column exists in the table, and then remove the column if previous statement is true. The database I use is Sybase ASE, and this is the code that I tried to use:
IF EXISTS (SELECT 1 FROM syscolumns WHERE id = object_id('users') AND name = 'maiden_name')
BEGIN
ALTER TABLE security DROP maiden_name
END
The code above executed successfully first time I run it. The second time I goth the error:
Invalid column name 'maiden_name'
If column does not exist the ALTER TABLE block of code shouldn't run. Is there a way to achieve this is Sybase? Thank you.
You can use dynamic SQL:
IF EXISTS (SELECT 1 FROM syscolumns WHERE id = object_id('users') AND name = 'maiden_name')
BEGIN
EXEC('ALTER TABLE security DROP maiden_name')
END;
The problem is that the parser is trying to parse the ALTER during the compilation phase, and it gets an error if the column does not exist.
I'm having a strange problem with an MSSQL Query that I'm trying to run in Microsoft SQL Server 2014. It is an update script for my database structure. It should basically rename a Column (from Price to SellingPrice) of a Table after its content was merged to another one.
USE db_meta
GO
DECLARE #BakItemPrices TABLE
(
ItemNum int,
Price int,
CashPrice int
)
-- backup old prices
insert into #BakItemPrices
select ItemNum, Price from dbo.ItemInfo
-- merge into other table
alter table ShopInfo
add column Price int NOT NULL DEFAULT ((0))
update ShopInfo
set ShopInfo.Price = i.Price
from ShopInfo s
inner join #BakItemPrices i
on s.ItemNum = i.ItemNum
GO
-- rename the column
exec sp_rename 'ItemInfo.Price', 'SellingPrice', 'COLUMN' -- The Debugger executes this first
GO
This query always gave me the error
Msg 207, Level 16, State 1, Line 13
Invalid column name 'Price'.
I couldn't understand this error until I debugged the query. I was amazed as I saw that the debugger wont even hit the breakpoint I placed at the backup code and says that "its unreachable because another batch is being executed at the moment".
Looking further down I saw that the debugger instantly starts with the exec sp_rename ... line before it executes the query code that I wrote above. So at the point my backup code is being executed the Column is named SellingPrice and not Price which obviously causes it to fail.
I thought queries get processed from top to bottom? Why is the execute sequence being executed before the code that I wrote above?
Script is sequenced from top to down. But some changes to schema is "visible" after the transaction with script is committed. Split your script into two scripts, it can help.
Does anyone see what's wrong with this code for SQL Server?
IF NOT EXISTS(SELECT *
FROM sys.columns
WHERE Name = 'OPT_LOCK'
AND object_ID = Object_id('REP_DSGN_SEC_GRP_LNK'))
BEGIN
ALTER TABLE REP_DSGN_SEC_GRP_LNK
ADD OPT_LOCK NUMERIC(10, 0)
UPDATE REP_DSGN_SEC_GRP_LNK
SET OPT_LOCK = 0
ALTER TABLE REP_DSGN_SEC_GRP_LNK
ALTER COLUMN OPT_LOCK NUMERIC(10, 0) NOT NULL
END;
When I run this, I get:
Msg 207, Level 16, State 1, Line 3
Invalid column name 'OPT_LOCK'.
on the update command.
Thanks.
In this case you can avoid the problem by adding the column as NOT NULL and setting the values for existing rows in one statement as per my answer here.
More generally the problem is a parse/compile issue. SQL Server tries to compile all statements in the batch before executing any of the statements.
When a statement references a table that doesn't exist at all the statement is subject to deferred compilation. When the table already exists it throws an error if you reference a non existing column. The best way round this is to do the DDL in a different batch from the DML.
If a statement both references a non existing column in an existing table and a non existent table the error may or may not be thrown before compilation is deferred.
You can either submit it in separate batches (e.g. by using the batch separator GO in the client tools) or perform it in a child scope that is compiled separately by using EXEC or EXEC sp_executesql.
The first approach would require you to refactor your code as an IF ... cannot span batches.
IF NOT EXISTS(SELECT *
FROM sys.columns
WHERE Name = 'OPT_LOCK'
AND object_ID = Object_id('REP_DSGN_SEC_GRP_LNK'))
BEGIN
ALTER TABLE REP_DSGN_SEC_GRP_LNK
ADD OPT_LOCK NUMERIC(10, 0)
EXEC('UPDATE REP_DSGN_SEC_GRP_LNK SET OPT_LOCK = 0');
ALTER TABLE REP_DSGN_SEC_GRP_LNK
ALTER COLUMN OPT_LOCK NUMERIC(10, 0) NOT NULL
END;
The root cause of the error is the newly added column name is not reflected in the sys.syscolumns and sys.columns table until you restart SQL Server Management Studio.
For your information,you can replace the IF NOT EXISTS with the COL_LENGTH function. It takes two parameters,
Table Name and
Column you are searching for
If the Column is found then it returns the range of the datatype of the column Ex: Int (4 bytes), when not found then it returns a NULL.
So, you could use this as follows and also combine 3 Statements into one.
IF (SELECT COL_LENGTH('REP_DSGN_SEC_GRP_LNK','OPT_LOCK')) IS NULL
BEGIN
ALTER TABLE REP_DSGN_SEC_GRP_LNK
ADD OPT_LOCK NUMERIC(10, 0) NOT NULL DEFAULT 0
END;
Makes it simpler.
Given the following SQL:
IF EXISTS (SELECT * FROM sys.columns WHERE name = 'NewFieldName' AND object_id = OBJECT_ID('dbo.MyTableName'))
RETURN
-- Add NewFieldName column to part of the Summer 2012 release cycle.
ALTER TABLE dbo.[MyTableName] ADD
[NewFieldName] SmallINT NOT NULL
CONSTRAINT DF_MyTableName_NewFieldName DEFAULT (2)
UPDATE [MyTableName] SET NewFieldName = 1 WHERE [Name] = 'FindMe' --Update one specific value
Produces the following error message:
Msg 207, Level 16, State 1, Line 10 Invalid column name
'NewFieldName'.
I'm sure I'm missing something basic, but trying to put "GO" after the alter makes the UPDATE run everytime and I don't want to do that.
How can I structure this statement so that it will check to see if the column exists and, if it doesn't add it and then set the values as stated in my UPDATE statements?
You need the statement referencing the new column to be compiled after the new column is added. One way of doing this is to run it as a child batch with EXEC.
IF NOT EXISTS (SELECT *
FROM sys.columns
WHERE name = 'NewFieldName'
AND object_id = OBJECT_ID('dbo.MyTableName'))
BEGIN
-- Add NewFieldName column to part of the Summer 2012 release cycle.
ALTER TABLE dbo.[MyTableName]
ADD [NewFieldName] SMALLINT NOT NULL
CONSTRAINT DF_MyTableName_NewFieldName DEFAULT (2)
EXEC(' UPDATE [MyTableName] SET NewFieldName = 1 WHERE [Name] = ''FindMe''')
END
The reason it worked for you originally is presumably because the table itself did not exist when the batch was compiled thus meaning that all statements in it referencing the table are subject to deferred compile.
I've written a stored procedure as following:
CREATE PROC spSoNguoiThan
#SNT int
AS
begin
IF not exists (select column_name from INFORMATION_SCHEMA.columns where
table_name = 'NhanVien' and column_name = 'SoNguoiThan')
ALTER TABLE NhanVien ADD SoNguoiThan int
else
begin
UPDATE NhanVien
SET NhanVien.SoNguoiThan = (SELECT Count(MaNguoiThan)FROM NguoiThan
WHERE MaNV=NhanVien.MaNV
GROUP BY NhanVien.MaNV)
end
SELECT *
FROM NhanVien
WHERE SoNguoiThan>#SNT
end
GO
Then I get the error :
Server: Msg 207, Level 16, State 1, Procedure spSoNguoiThan, Line 12
Invalid column name 'SoNguoiThan'.
Server: Msg 207, Level 16, State 1, Procedure spSoNguoiThan, Line 15
Invalid column name 'SoNguoiThan'.
Who can help me?
Thanks!
When the stored proc is parsed during CREATE the column does not exist so you get an error.
Running the internal code line by line works because they are separate. The 2nd batch (UPDATE) runs because the column exists.
The only way around this would be to use dynamic SQL for the update and select so it's not parsed until EXECUTE time (not CREATE time like now).
However, this is something I really would not do: DDL and DML in the same bit of code
I ran into this same issue and found that in addition to using dynamic sql I could solve it by cross joining to a temp table that had only one row. That caused the script compiler to not try to resolve the renamed column at compile time. Below is an example of what I did to solve the issue without using dynamic SQL
select '1' as SomeText into #dummytable
update q set q.ValueTXT = convert(varchar(255), q.ValueTXTTMP) from [dbo].[SomeImportantTable] q cross join #dummytable p