|
database
newsgroups
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Check ConstraintDateTime field is used in the unique value, but I only need to the DatePart (not the time), otherwise I would use a standard unique constraint. I need the combination of ResID,RxDayID,RxTimeID, and Just The Date of DispDateTime to be unique. The following works for inserts but not updates if a record already exists. It does work if there are 2 matching records for an update. I'd rather have the back-end enforce the rules, but if I must I'll build it into the application. Is there a function or a method I'm over-looking? ALTER TABLE [dbo].[ResRxDispensed] ADD CONSTRAINT [CK_ResRxDispensed_Unique] CHECK ([dbo].[DispRecordUnique_FN]( [ResID], [RxDayID], [RxDoseTimeID], [DispDateTime]) = 'Y') Create Function [dbo].[DispRecordUnique_FN]( @ResID Int, @RxDayID Int, @RxTimeID Int, @ADate DateTime) Returns VarChar(1) Begin Declare @TheDate DateTime, @Rtn Varchar(1),@Cnt Int Set @TheDate = Cast(Cast(@ADate as Char(11)) As DateTime) Set @Rtn = 'N' Select @Cnt = Count(*) From ResRxDispensed Where ResID = @ResID And RxDayID = @RxDayID And RxDoseTimeID = @RxTimeID And (DispDateTime >= @TheDate And DispDateTime <= (@TheDate+1)) If (@Cnt <= 1) Begin Set @Rtn = 'Y' End Return @Rtn End TIA, -Steve- You've overcomplicated the solution IMHO:
1) create a computed column: alter table dbo.ResRxDispensed add DispDate as convert(char(8), DispDateTime, 112) go 2) create the unique constraint: alter table dbo.ResRxDispensed add constraint <constraint name> unique (ResID, RxDayID, RxTimeID, DispDate) go If this does not help, please post proper DDL and maybe sample data. ML --- http://milambda.blogspot.com/ Thanks.
I tried your suggestion but it apparently doesn't work when you use a computed column in a constraint. When I attempt to add a record I get the errror: [INSERT failed because the following SET options have incorrect settings: 'ARITHABORT'] Using SET ARITHABORT ON before the insert/update works, but it seems it must be issued before each update/insert statement. > If this does not help, please post proper DDL and maybe sample data. CREATE TABLE [dbo].[ResRxDispensed] ([ID] [int] IDENTITY (1, 1) NOT NULL , [ResID] [int] NULL , [ResRxID] [int] NULL , [RxDayID] [int] NULL , [RxDoseTimeID] [int] NULL , [DispDateTime] [datetime] NULL , [Created] [datetime] NULL , [StaffID] [int] NULL , [DispDate] AS (convert(char(8),[DispDateTime],112)) ) ON [PRIMARY] Insert Into ResRxDispensed ([ResID],[ResRxID],[RxDayID],[RxDoseTimeID],[DispDateTime]) Values (1,1,1,1,{ts '2006-04-06 09:30:00'}) ; The insert should fail if another record has ResID = 1 RxDayID = 1 RxDoseTimeID = 1 DispDateTime = 2006-04-06 (with any time factor) Updates should also fail if the update changes the record to match another record with the criteria. -Steve- This worked fine for me on SQL 2000.
CREATE TABLE [dbo].[TestTable] ( [ResID] [char] (10) COLLATE SQL_Latin1_General_CP1_CI_AS NOT NULL , [RxDayID] [char] (10) COLLATE SQL_Latin1_General_CP1_CI_AS NOT NULL , [RxDoseTimeID] [char] (10) COLLATE SQL_Latin1_General_CP1_CI_AS NOT NULL , [DispDateTime] [datetime] NOT NULL , [OtherData] [varchar] (50) COLLATE SQL_Latin1_General_CP1_CI_AS NULL , [DispDate] AS (convert(char(8),[DispDateTime],112)) ) ON [PRIMARY] GO ALTER TABLE [dbo].[TestTable] ADD CONSTRAINT [IX_TestTable] UNIQUE NONCLUSTERED ( [ResID], [RxDoseTimeID], [RxDayID], [DispDate] ) ON [PRIMARY] GO Show quote "Steve Zimmelman" <skz@charter.nospam.net> wrote in message news:ep7wCVaWGHA.1220@TK2MSFTNGP02.phx.gbl... > Thanks. > > I tried your suggestion but it apparently doesn't work when you use a computed > column in a constraint. When I attempt to add a record I get the errror: > [INSERT failed because the following SET options have incorrect settings: > 'ARITHABORT'] > > Using SET ARITHABORT ON before the insert/update works, but it seems it must be > issued before each update/insert statement. > > > If this does not help, please post proper DDL and maybe sample data. > > CREATE TABLE [dbo].[ResRxDispensed] ( > [ID] [int] IDENTITY (1, 1) NOT NULL , > [ResID] [int] NULL , > [ResRxID] [int] NULL , > [RxDayID] [int] NULL , > [RxDoseTimeID] [int] NULL , > [DispDateTime] [datetime] NULL , > [Created] [datetime] NULL , > [StaffID] [int] NULL , > [DispDate] AS (convert(char(8),[DispDateTime],112)) > ) ON [PRIMARY] > > > Insert Into ResRxDispensed > ([ResID],[ResRxID],[RxDayID],[RxDoseTimeID],[DispDateTime]) > Values > (1,1,1,1,{ts '2006-04-06 09:30:00'}) ; > > The insert should fail if another record has > ResID = 1 > RxDayID = 1 > RxDoseTimeID = 1 > DispDateTime = 2006-04-06 (with any time factor) > > Updates should also fail if the update changes the record to match another > record with the criteria. > > -Steve- > > > > This worked fine for me on SQL 2000. Hi Jim,Thanks. I'm using SQL 2000, but I'm still getting the same error. Here's the scripts for creating everything I'm using. CREATE TABLE [dbo].[ResRxDispensed] ( [ID] [int] IDENTITY (1, 1) NOT NULL , [ResID] [int] NULL , [ResRxID] [int] NULL , [RxDayID] [int] NULL , [RxDoseTimeID] [int] NULL , [DispDateTime] [datetime] NOT NULL , [Created] [datetime] NULL , [StaffID] [int] NULL , [DispDate] AS (convert(char(8),[DispDateTime],112)) ) ON [PRIMARY] GO ALTER TABLE [dbo].[ResRxDispensed] WITH NOCHECK ADD CONSTRAINT [PK_ResRxDispensed] PRIMARY KEY CLUSTERED ( [ID] ) ON [PRIMARY] GO ALTER TABLE [dbo].[ResRxDispensed] ADD CONSTRAINT [DF_ResRxDispensed_DispDateTime] DEFAULT (getdate()) FOR [DispDateTime], CONSTRAINT [DF_ResRxDispensed_Created] DEFAULT (getdate()) FOR [Created], CONSTRAINT [IX_ResRxDispensed] UNIQUE NONCLUSTERED ( [ResID], [RxDayID], [RxDoseTimeID], [DispDate] ) ON [PRIMARY] GO CREATE INDEX [ResRxDispensed_ResID] ON [dbo].[ResRxDispensed]([ResID]) ON [PRIMARY] GO CREATE INDEX [ResRxDispensed_ResRxID] ON [dbo].[ResRxDispensed]([ResRxID]) ON [PRIMARY] GO CREATE INDEX [ResRxDispensed_RxDayID] ON [dbo].[ResRxDispensed]([RxDayID]) ON [PRIMARY] GO CREATE INDEX [ResRxDispensed_RxDoseTimeID] ON [dbo].[ResRxDispensed]([RxDoseTimeID]) ON [PRIMARY] GO -Steve- Is there a reason why you can't add this SET option to your stored procedures?
-- Show quoteTibor Karaszi, SQL Server MVP http://www.karaszi.com/sqlserver/default.asp http://www.solidqualitylearning.com/ Blog: http://solidqualitylearning.com/blogs/tibor/ "Steve Zimmelman" <skz@charter.nospam.net> wrote in message news:%23Set$paWGHA.2080@TK2MSFTNGP05.phx.gbl... >> This worked fine for me on SQL 2000. > > Hi Jim, > > Thanks. I'm using SQL 2000, but I'm still getting the same error. > > Here's the scripts for creating everything I'm using. > > CREATE TABLE [dbo].[ResRxDispensed] ( > [ID] [int] IDENTITY (1, 1) NOT NULL , > [ResID] [int] NULL , > [ResRxID] [int] NULL , > [RxDayID] [int] NULL , > [RxDoseTimeID] [int] NULL , > [DispDateTime] [datetime] NOT NULL , > [Created] [datetime] NULL , > [StaffID] [int] NULL , > [DispDate] AS (convert(char(8),[DispDateTime],112)) > ) ON [PRIMARY] > GO > > ALTER TABLE [dbo].[ResRxDispensed] WITH NOCHECK ADD > CONSTRAINT [PK_ResRxDispensed] PRIMARY KEY CLUSTERED > ( > [ID] > ) ON [PRIMARY] > GO > > ALTER TABLE [dbo].[ResRxDispensed] ADD > CONSTRAINT [DF_ResRxDispensed_DispDateTime] DEFAULT (getdate()) FOR [DispDateTime], > CONSTRAINT [DF_ResRxDispensed_Created] DEFAULT (getdate()) FOR [Created], > CONSTRAINT [IX_ResRxDispensed] UNIQUE NONCLUSTERED > ( > [ResID], > [RxDayID], > [RxDoseTimeID], > [DispDate] > ) ON [PRIMARY] > GO > > CREATE INDEX [ResRxDispensed_ResID] ON [dbo].[ResRxDispensed]([ResID]) ON [PRIMARY] > GO > > CREATE INDEX [ResRxDispensed_ResRxID] ON [dbo].[ResRxDispensed]([ResRxID]) ON [PRIMARY] > GO > > CREATE INDEX [ResRxDispensed_RxDayID] ON [dbo].[ResRxDispensed]([RxDayID]) ON [PRIMARY] > GO > > CREATE INDEX [ResRxDispensed_RxDoseTimeID] ON [dbo].[ResRxDispensed]([RxDoseTimeID]) ON [PRIMARY] > GO > > > -Steve- > "Tibor Karaszi" wrote
> Is there a reason why you can't add this SET option to your stored procedures? I suppose not. I just wanted something that didn't require such special handling for updates in case I needed to do some manual repair/entry outside of the application. -Steve- "Tibor Karaszi" wrote: This procedure, when used, produces the same error.> Is there a reason why you can't add this SET option to your stored procedures? [INSERT failed because the following SET options have incorrect settings: 'ARITHABORT'] I'm at a loss how to proceed... -Steve- Exec NewResRxDispensed_SP 1,1,1,1,{ts '2006-05-13 09:30:00'},24 Create Procedure [dbo].[NewResRxDispensed_SP] @ResID int, @ResRxID int, @RxDayID int, @RxDoseTimeID int, @DispDateTime datetime, @StaffID int As SET ARITHABORT ON Insert Into [ResRxDispensed] ([ResID], [ResRxID], [RxDayID], [RxDoseTimeID], [DispDateTime], [StaffID]) Values (@ResID, @ResRxID, @RxDayID, @RxDoseTimeID, @DispDateTime, @StaffID) /*** Return New Int ID [ID] ***/ Select SCOPE_IDENTITY() As NewID Consider using ARITHABORT ON as general database option. But read up on
possible issues in Books Online. ML --- http://milambda.blogspot.com/ That is strange. I executed the DDL posted a few posts ago, and then created the procedure you
posted and then executed the procedure. No error message. I did this on both SQL2K sp3 and SQL2005. Perhaps you can try the same (in tempdb or a new database) just to rule out other factors? -- Show quoteTibor Karaszi, SQL Server MVP http://www.karaszi.com/sqlserver/default.asp http://www.solidqualitylearning.com/ Blog: http://solidqualitylearning.com/blogs/tibor/ "Steve Zimmelman" <skz@charter.nospam.net> wrote in message news:eGzuDCbWGHA.1196@TK2MSFTNGP03.phx.gbl... > "Tibor Karaszi" wrote: >> Is there a reason why you can't add this SET option to your stored procedures? > > This procedure, when used, produces the same error. > [INSERT failed because the following SET options have incorrect settings: 'ARITHABORT'] > > I'm at a loss how to proceed... > > -Steve- > > Exec NewResRxDispensed_SP 1,1,1,1,{ts '2006-05-13 09:30:00'},24 > > Create Procedure [dbo].[NewResRxDispensed_SP] > @ResID int, > @ResRxID int, > @RxDayID int, > @RxDoseTimeID int, > @DispDateTime datetime, > @StaffID int > As > SET ARITHABORT ON > Insert Into [ResRxDispensed] > ([ResID], > [ResRxID], > [RxDayID], > [RxDoseTimeID], > [DispDateTime], > [StaffID]) > Values > (@ResID, > @ResRxID, > @RxDayID, > @RxDoseTimeID, > @DispDateTime, > @StaffID) > /*** Return New Int ID [ID] ***/ > Select SCOPE_IDENTITY() As NewID > > > > I finally gave up and wrote a trigger for insert and updates. It works this
way, so I'll leave it for now and move on. Thanks everyone. I appreciate your time and advise. -Steve- Show quote "Tibor Karaszi" <tibor_please.no.email_kara***@hotmail.nomail.com> wrote in message news:%23bq0wnhWGHA.4324@TK2MSFTNGP03.phx.gbl... > That is strange. I executed the DDL posted a few posts ago, and then created > the procedure you posted and then executed the procedure. No error message. I > did this on both SQL2K sp3 and SQL2005. Perhaps you can try the same (in > tempdb or a new database) just to rule out other factors? > > -- > Tibor Karaszi, SQL Server MVP Experiment to see if the year(), month(), day() or datediff() functions
would work here. Show quote "Steve Zimmelman" <skz@charter.nospam.net> wrote in message news:%23kFjHVZWGHA.2180@TK2MSFTNGP02.phx.gbl... > I'm having a little trouble enforcing a unique record rule. The problem > is a DateTime field is used in the unique value, but I only need to the > DatePart (not the time), otherwise I would use a standard unique > constraint. > > I need the combination of ResID,RxDayID,RxTimeID, and Just The Date of > DispDateTime to be unique. > > The following works for inserts but not updates if a record already > exists. It does work if there are 2 matching records for an update. I'd > rather have the back-end enforce the rules, but if I must I'll build it > into the application. Is there a function or a method I'm over-looking? > > ALTER TABLE [dbo].[ResRxDispensed] ADD > CONSTRAINT [CK_ResRxDispensed_Unique] > CHECK ([dbo].[DispRecordUnique_FN]( > [ResID], > [RxDayID], > [RxDoseTimeID], > [DispDateTime]) = 'Y') > > > Create Function [dbo].[DispRecordUnique_FN]( > @ResID Int, > @RxDayID Int, > @RxTimeID Int, > @ADate DateTime) > Returns VarChar(1) > Begin > Declare @TheDate DateTime, @Rtn Varchar(1),@Cnt Int > Set @TheDate = Cast(Cast(@ADate as Char(11)) As DateTime) > Set @Rtn = 'N' > Select @Cnt = Count(*) > From ResRxDispensed > Where ResID = @ResID > And RxDayID = @RxDayID > And RxDoseTimeID = @RxTimeID > And (DispDateTime >= @TheDate And DispDateTime <= (@TheDate+1)) > If (@Cnt <= 1) > Begin > Set @Rtn = 'Y' > End > Return @Rtn > End > > TIA, > -Steve- > "JT" <some***@microsoft.com> wrote in message I'm not sure I understand. Experiment how? The function news:%23sosFSaWGHA.4424@TK2MSFTNGP05.phx.gbl... > Experiment to see if the year(), month(), day() or datediff() functions would > work here. DispRecordUnique_FN()works, but using it in a Check constraint only works on Inserts, not updates. -Steve- |
|||||||||||||||||||||||