Description: String cannot have zero length. Event ID: 6003

When i add a new document to the queue, I receive this error in the event log. I used the example from the book in chapter 5 but i left out the scheduled piece. I am unable to find anything via searches on google etc. Your help would be appreciated. Thank you!

XSLT File

<?xml version="1.0" encoding="UTF-8" ?>

<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">

<xsl:output method="text"/>

<xsl:template match="notifications">

There is a new Test Document!

<xsl:apply-templates/>

</xsl:template>

<xsl:template match="notification">

Document Type: <xsl:value-of select="TypeName" />

Document Name: <xsl:value-of select="DocumentTitle" />

Document Category: <xsl:value-of select="CategoryName" />

<xsl:text>

</xsl:text>

</xsl:template>

</xsl:stylesheet>

Event log error:

Description: The assembly containing the content formatter could not be loaded.

EventParameters:

Instance Name: MyInstance 'Renamed for forum'

Application Name: Documents

Content Formatter Assembly Name:

Description: String cannot have zero length.

InstanceName: MyInstance 'Renamed for forum'

ApplicationName: Documents

Component: Distributor

ComponentDetails: DistributorName: MyDistributor 'Renamed for forum'

Thread: 4

ACF:

<ContentFormatter>

<ClassName>XsltFormatter</ClassName>

<Arguments>

<Argument>

<Name>XsltBaseDirectoryPath</Name>

<Value>%_ApplicationBaseDirectoryPath_%\Formatters</Value>

</Argument>

<Argument>

<Name>XsltFileName</Name>

<Value>NewDocumentByType.xslt</Value>

</Argument>

</Arguments>

</ContentFormatter>

ICF

<Application>

<ApplicationName>Documents</ApplicationName>

<BaseDirectoryPath>%_InstanceBaseDirectoryPath_%\Documents</BaseDirectoryPath>

<ApplicationDefinitionFilePath>ApplicationDefinition.xml</ApplicationDefinitionFilePath>

<Parameters>

<Parameter>

<Name>_NSServer_</Name>

<Value>%_NSServer_%</Value>

</Parameter>

<Parameter>

<Name>_SQLServer_</Name>

<Value>%_SQLServer_%</Value>

</Parameter>

<Parameter>

<Name>_ApplicationBaseDirectoryPath_</Name>

<Value>%_InstanceBaseDirectoryPath_%\Documents</Value>

</Parameter>

</Parameters>

</Application>

</Applications>

I fixed this by building a custom Formatter.

Derived Tables

OK...I know how to write a query to return for example :

All the people that ordered X and Y

but how do I write one for:

All the people that ordered X but not Y?

Thanks,
TreyO yea...this is how i did the first part

SELECT DISTINCT c.Company
FROM Customers as c
JOIN
 (SELECT CustomerID
 FROM Orders o
 JOIN [Order Details] od
 ON o.OrderID = od.OrderID
 JOIN products p
 ON od.ProductID = p.ProductID
 WHERE p.ProductName = 'X') as temp1
 ON c.CustomerID = temp1.CustomerID
JOIN
 (SELECT CustomerID
 FROM Orders o
 JOIN [Order Details] od
 ON o.OrderID = od.OrderID
 JOIN products p
 ON od.ProductID = p.ProductID
 WHERE p.ProductName = 'Y') as temp2
 ON c.CustomerID = temp2.CustomerID

Derived Column Task Replace Quotes

How do you replace quotes in an expression

I mean for example if I needed to replace xx in a string with empty string then the following works: REPLACE(SelectedString, "xx","")

But the example I have needs to actually replace quote marks in a string with an empty string and REPLACE(SelectedString, " " ","") doesn't work. I tried guessing a few option like &QUOTE or &QTE or something...

Any ideas ?

Thanks

Richard

Got it !! :

REPLACE([Selected Closing],"\"","")

Derived Column Calculation

Example: (47 / 204709) * 12000 = 2.754

Both values (47 & 204709) are of data type Int.

SQL Sever - Derived Column Calculation returns 2.74

The destination data type is Float

I have converted the data types to Float, Decimal & Numeric and still got the same answer.

Any suggestions

Have you casted the result of the calculation to a float?

-Jamie

|||

Yes, I have what I have noticed with the Data Convertion component is that the input and output data types stay the same. Should it not change to the new data type.

|||

Can you post your expression up here?

-Jamie

|||

Michael Exall wrote:

Yes, I have what I have noticed with the Data Convertion component is that the input and output data types stay the same. Should it not change to the new data type.

No. If you were changing the data-type of a column then inherently you are changing the "shape" of the memory buffer and therefore this would need to be an asynchronous component.

A design choice was obviously made to keep data conversion component synchronous for perf reasons.

If you're puzzled as to the difference between synch and asynch...google it http://www.google.co.uk/search?hl=en&q=ssis+synchronous+asynchronous&meta=

-Jamie

Derived Column - Redirect Row

Currently have a derived column as part of a dataflow to help combat spaces
and blanks when converting data types.
example:
derived column 1: (DT_I4)[DUEQTY] if the column can't be converted it
redirects the row then another derived column process uses (DT_I4)0
I have this concept on three columns within the table.
if there are only errors on one of the columns it redirects the whole row
and changes all values to zero.
How do I combat this? I think i'm looking for a derived column variable
statment that is similiatr to the if(iserror(statement),"",(statement)) in
Excel
--
Milk BottleHi
Is it SQL Server's statement? Can you show us?
"MilkBottle" <MilkBottle@.discussions.microsoft.com> wrote in message
news:57051AEA-0342-4A9B-B67F-E42FB34B0036@.microsoft.com...
> Currently have a derived column as part of a dataflow to help combat
> spaces
> and blanks when converting data types.
> example:
> derived column 1: (DT_I4)[DUEQTY] if the column can't be converted it
> redirects the row then another derived column process uses (DT_I4)0
> I have this concept on three columns within the table.
> if there are only errors on one of the columns it redirects the whole row
> and changes all values to zero.
> How do I combat this? I think i'm looking for a derived column variable
> statment that is similiatr to the if(iserror(statement),"",(statement)) in
> Excel
> --
> Milk Bottle|||> if there are only errors on one of the columns it redirects the whole row
> and changes all values to zero.
Rather than set all values to zero, set only the problem column (e.g.
column1) to zero. Connect the outputs of both column1 derrived columns to a
Union All so that both good and fixed records continue down the pipeline.
Repeat this for the remaining columns you want to scrub.
Hope this helps.
Dan Guzman
SQL Server MVP
"MilkBottle" <MilkBottle@.discussions.microsoft.com> wrote in message
news:57051AEA-0342-4A9B-B67F-E42FB34B0036@.microsoft.com...
> Currently have a derived column as part of a dataflow to help combat
> spaces
> and blanks when converting data types.
> example:
> derived column 1: (DT_I4)[DUEQTY] if the column can't be converted it
> redirects the row then another derived column process uses (DT_I4)0
> I have this concept on three columns within the table.
> if there are only errors on one of the columns it redirects the whole row
> and changes all values to zero.
> How do I combat this? I think i'm looking for a derived column variable
> statment that is similiatr to the if(iserror(statement),"",(statement)) in
> Excel
> --
> Milk Bottle

Deployment 'utility' script using sqlcmd

Hi all,

I'm looking at creating a sample utility script that will invoking

scripts to deploy some SQL code. For example, a utlity script that will

run a SQL script, and on successful completion, execute the next script.

Having not used SQLCMD at all before, and being very new to SQL2005

(< 1 month) please guide me if there is a better way of invoking

this... For example, a way of avoiding the xp_cmdshell invocation!

The following code invokes a script, but I'm trying to find a way of

getting a return code back from sqlcmd, so I can progress and do the

next, or fail if the return code <> 0 (success).

[code]

--Process to create DB, Tables, and Stored Procedures

set nocount on

DECLARE

@.Error int,

@.ExecCommand varchar(512),

@.FullFilePath varchar(255)

--create the database

BEGIN TRY

SET @.FullFilePath =

'D:\Documentation\Projects\Integration Services\BIDS Projects\Tesco DNF

Integration Services\TescoDNF ProductPromo\SQL Code\OBJECTS\Create DB

TescoDNF_SSISPackageManager.sql'

SET @.ExecCommand = 'xp_cmdshell ''sqlcmd -S Rgalbraith\SQL2005_1 -i "'+@.FullFilePath+'"'' '

SELECT @.FullFilePath

UNION

SELECT @.ExecCommand

EXEC (@.ExecCommand)

SELECT @.@.ERROR

SELECT @.Error

END TRY

BEGIN CATCH

SELECT

ERROR_NUMBER() AS ErrorNumber,

ERROR_SEVERITY() AS ErrorSeverity,

ERROR_STATE() AS ErrorState,

ERROR_PROCEDURE() AS ErrorProcedure,

ERROR_LINE() AS ErrorLine,

ERROR_MESSAGE() AS ErrorMessage;

GOTO ErrorAbort

END CATCH
ErrorAbort:

[/code]

Hello
I didn't quite understand what are you trying to achieve. You have some sequence of sql scripts that need to be executed against a server one after another, if no error occurs?
Then, what if an error occurs? Maybe there is some branching in the scripts? I.e. if script1.sql succeedes, then execute script2.sql, else execute script3.sql. If script3.sql fails, restore a database backup...
And why are you doing this from SQL? Isn't using a programming language more effective?

|||My thought had been to have a simple T-SQL script that 'deploys' a set

of SQL scripts to, for example, a server. For example, some pseudo-code

Create Database

If Error abort

Create Table1

If Error abort

Create Table2

If Error Abort

Create Stored Procedure1

If Error Abort

ELSE Complete and report success

I do agree that this is something that could be (better) done in a

"proper" coding language like .Net, c# etc. but (a) it's just a simple

utility script (b) it teaches me moore about usage of SQLCMD and (c) I

do not have any skill in any normal programming language, hence I was

planning to write a quick deployment utility with a script.

The idea might be something as ugly as a table structure that has a

list of scripts registered in it, with some sequence logic - like

creating parent tables before children tables - and then a cursor (or a

better method if I can find it) that fetches a sqlcmd filename

execution command, executes it, and on success fetches the next one

based on the sequence logic.

I can probably do all of that in about 4 hours in T-SQL, if I can find

a way to confirm the successful execution of the previous command....|||

Ok I got it.
You can have a "Version" table, then number your scripts so that each of them updates the version. Before executing each portion of code, you can check the current version to be exactly the number you need.
e.g.
Create Database
Create table Version(VersionNum varchar(255), ChangeID int)
insert into Version(VersionNum, ChanegeID) values("1", @.ID)
GO
If (select VersionNum where ChangeID=@.ID)="1"
BEGIN
Create Table1
Update Version set VersionNum = "2" where ChangeID=@.ID
END
GO
If (select VersionNum where ChangeID=@.ID)="2"
...
So, basicly you just update the version number as the last command of each batch. Then, you check for the appropriate version number at the beginning of the next block.
This way you can even do some "branching". Even more, if your scripts fails, you can check what scripts have succeeded and what scripts have not, simply by looking at the VersionNo field.

Anyhow, I'd strongly reccomend using ordinal programming language if you are going to use that utility more than once and it MIGHT become somehow complicated.

|||Well, in a sense. The point is though that I want to fetch sql files

and execute them, and not merge them all into a single large script.

So, I want utility script to do this:

Run external sql script

On failure abort, on Success

Run external sql script

On Failure abort, On Success

...

You being to see why I referred to a cursor?

The point is that the utility script wouldn't contain any of the client

SQL commands - it would fetch them by referring to the table, and

fetching the path to the SQL file, and building a SQLCMD to execute

that script

I guess, as you say, I could add a generic update ##SQLScriptTracker

table, then check it on the new execution, or abort. I had hoped for a

neater solution - i.e. SQLCMD being able to return a returncode that it

gets from a SQL file it ran....|||

Why dont go for Batch files (.bat). There you can execute the individual script files one by one using the SQLCMD. And for aborting when error occurs, check the ERRORLEVEL, if its not 0 then quit execution or skip to other location using GOTO.

echo Backup database
sqlcmd -S(local) -U<uid> -P<pwd> -i"backup_db.sql"
IF ERRORLEVEL 1 GOTO abort_bkp

echo Update database
sqlcmd -S(local) -U<uid> -P<pwd> -i"create_proc.sql"
IF ERRORLEVEL 1 GOTO abort

echo Update customer data
sqlcmd -S(local) -U<uid> -P<pwd> -i"update_customer_data.sql"
IF ERRORLEVEL 1 GOTO abort_with_restore

:abort_bkp
echo Error backup database. Setup aborted

:abort_with_restore
echo Error updating data. Restoring database...
sqlcmd -S(local) -U<uid> -P<pwd> -i"restore_db.sql"
IF ERRORLEVEL 1 GOTO res_falied
...
...

|||

hmm ... it seems as thought ERRORLEVEL is only set on the SUCCESS/FAILURE of the SQLCMD invocation, and not based on the SUCCESS/FAILURE of the invoked sql commands?

for example:

batch CALLBACKUP.BAT file contents:

echo Backup database
sqlcmd -SRGalbraith\SQL2005_1 -E -i"d:\backup_db.sql"
IF ERRORLEVEL 1 GOTO abort_bkp
IF ERRORLEVEL 0 GOTO done

:abort_bkp
echo Error backup database. Setup aborted

:done
echo all done now

backup_db.sql contents

backup database DataStore2 to disk = 'D:\BackupDatabase.bak'

execution results:

D:\>sqlcmd -SRGalbraith\SQL2005_1 -E -i"d:\backup_db.sql"
Msg 911, Level 16, State 11, Server RGALBRAITH\SQL2005_1, Line 1
Could not locate entry in sysdatabases for database 'DataStore2'. No entry found with that name. Make sure that the name
is entered correctly.
Msg 3013, Level 16, State 1, Server RGALBRAITH\SQL2005_1, Line 1
BACKUP DATABASE is terminating abnormally.

D:\>IF ERRORLEVEL 1 GOTO abort_bkp

D:\>IF ERRORLEVEL 0 GOTO done

D:\>echo all done now
all done now

A sample of sqlcmd failing was:

D:\>callbackup

D:\>echo Backup database
Backup database

D:\>sqlcmd -SRGalbraith\SQL2005_1 -E -i"d:\backup_db.sql"
Sqlcmd: 'd:\backup_db.sql': Invalid filename.

D:\>IF ERRORLEVEL 1 GOTO abort_bkp

D:\>echo Error backup database. Setup aborted
Error backup database. Setup aborted

D:\>echo all done now
all done now

...

As is probably obvious, I'm not much of a batch file coder :-), but the jmist of it is there - when the SQLCMD failed (file not found) then it reported error, but when the SQL script failed (database not found) no error was reported. Is there a way around that?

|||

You have to set the -b option for the SQLCMD. -b makes the batch abort with an error if the script fails. So you would write this...

@.ECHO OFF

@.echo.
@.echo Backup database
sqlcmd -S.\sqlexpress -E -i"backup_db.sql" -b
IF %ERRORLEVEL% NEQ 0 GOTO err_bkp_failed

:success
echo Database update successful
goto end

:err_bkp_failed
echo Backup failed. Aborting...
goto end

:end

HTH

|||hmm - good to know! still going to investiage the other options as

well, since with the batch file I have to add a file each time.

Thanks|||

Visual Studio .Net 2003 had a "create batch file" command which was beautiful for creating this batch file to process the sequence of sql scripts that you create.

I still use it today. But it seems we are in need to migrate to Visual Studio 2005, and this feature has been disabled now.

Do you have a more elegant solution now?

|||

Actually you don't have to modifiy the bat script each time. I have been using bat scripts to do exactly this for years.

The shell support the For Each looping structure which will set a shell variable to each file name that meet's a spec.

For Each %%1 in *.sql <execute a dos command>

I have been using the OSQL command line utility for years like this. I guess I will have to update to SQLCMD now.

You can find out the details of shell commands by going to "My Computer" <Help> and searching for "For Each"

You can find out about OSQL in BOL

|||

I've been searching solution on catching MS SQL abortion errors in a launching batch file. With option '-b', at least the batch file could return error code 1 instead of 0. Thanks for the hint!

Still, I'd appreciate if anyone could offer answer on capturing the stdout error in the batch file. My problem is that once the sql statement is aborted, it immediately exits from the erroneous line, ignores the rest code in the same script. Therefore, no error could be saved.

Also, I found that in some env. the 'sqlcmd' is not recognized (SQL Server 2000?) but 'osql' or 'isql'. Are there any differences among them (must be, but I don't know).

Deployment 'utility' script using sqlcmd

Hi all,
I'm looking at creating a sample utility script that will invoking scripts to deploy some SQL code. For example, a utlity script that will run a SQL script, and on successful completion, execute the next script.
Having not used SQLCMD at all before, and being very new to SQL2005 (< 1 month) please guide me if there is a better way of invoking this... For example, a way of avoiding the xp_cmdshell invocation!
The following code invokes a script, but I'm trying to find a way of getting a return code back from sqlcmd, so I can progress and do the next, or fail if the return code <> 0 (success).
[code]
--Process to create DB, Tables, and Stored Procedures
set nocount on
DECLARE
@.Error int,
@.ExecCommand varchar(512),
@.FullFilePath varchar(255)
--create the database
BEGIN TRY
SET @.FullFilePath = 'D:\Documentation\Projects\Integration Services\BIDS Projects\Tesco DNF Integration Services\TescoDNF ProductPromo\SQL Code\OBJECTS\Create DB TescoDNF_SSISPackageManager.sql'
SET @.ExecCommand = 'xp_cmdshell ''sqlcmd -S Rgalbraith\SQL2005_1 -i "'+@.FullFilePath+'"'' '
SELECT @.FullFilePath
UNION
SELECT @.ExecCommand
EXEC (@.ExecCommand)
SELECT @.@.ERROR
SELECT @.Error
END TRY
BEGIN CATCH
SELECT
ERROR_NUMBER() AS ErrorNumber,
ERROR_SEVERITY() AS ErrorSeverity,
ERROR_STATE() AS ErrorState,
ERROR_PROCEDURE() AS ErrorProcedure,
ERROR_LINE() AS ErrorLine,
ERROR_MESSAGE() AS ErrorMessage;
GOTO ErrorAbort
END CATCH

ErrorAbort:
[/code]

Hello
I didn't quite understand what are you trying to achieve. You have some sequence of sql scripts that need to be executed against a server one after another, if no error occurs?
Then, what if an error occurs? Maybe there is some branching in the scripts? I.e. if script1.sql succeedes, then execute script2.sql, else execute script3.sql. If script3.sql fails, restore a database backup...
And why are you doing this from SQL? Isn't using a programming language more effective?

|||My thought had been to have a simple T-SQL script that 'deploys' a set of SQL scripts to, for example, a server. For example, some pseudo-code
Create Database
If Error abort
Create Table1
If Error abort
Create Table2
If Error Abort
Create Stored Procedure1
If Error Abort
ELSE Complete and report success
I do agree that this is something that could be (better) done in a "proper" coding language like .Net, c# etc. but (a) it's just a simple utility script (b) it teaches me moore about usage of SQLCMD and (c) I do not have any skill in any normal programming language, hence I was planning to write a quick deployment utility with a script.
The idea might be something as ugly as a table structure that has a list of scripts registered in it, with some sequence logic - like creating parent tables before children tables - and then a cursor (or a better method if I can find it) that fetches a sqlcmd filename execution command, executes it, and on success fetches the next one based on the sequence logic.
I can probably do all of that in about 4 hours in T-SQL, if I can find a way to confirm the successful execution of the previous command....
|||

Ok I got it.
You can have a "Version" table, then number your scripts so that each of them updates the version. Before executing each portion of code, you can check the current version to be exactly the number you need.
e.g.
Create Database
Create table Version(VersionNum varchar(255), ChangeID int)
insert into Version(VersionNum, ChanegeID) values("1", @.ID)
GO
If (select VersionNum where ChangeID=@.ID)="1"
BEGIN
Create Table1
Update Version set VersionNum = "2" where ChangeID=@.ID
END
GO
If (select VersionNum where ChangeID=@.ID)="2"
...
So, basicly you just update the version number as the last command of each batch. Then, you check for the appropriate version number at the beginning of the next block.
This way you can even do some "branching". Even more, if your scripts fails, you can check what scripts have succeeded and what scripts have not, simply by looking at the VersionNo field.

Anyhow, I'd strongly reccomend using ordinal programming language if you are going to use that utility more than once and it MIGHT become somehow complicated.

|||Well, in a sense. The point is though that I want to fetch sql files and execute them, and not merge them all into a single large script.
So, I want utility script to do this:
Run external sql script
On failure abort, on Success
Run external sql script
On Failure abort, On Success
...
You being to see why I referred to a cursor?
The point is that the utility script wouldn't contain any of the client SQL commands - it would fetch them by referring to the table, and fetching the path to the SQL file, and building a SQLCMD to execute that script
I guess, as you say, I could add a generic update ##SQLScriptTracker table, then check it on the new execution, or abort. I had hoped for a neater solution - i.e. SQLCMD being able to return a returncode that it gets from a SQL file it ran....
|||

Why dont go for Batch files (.bat). There you can execute the individual script files one by one using the SQLCMD. And for aborting when error occurs, check the ERRORLEVEL, if its not 0 then quit execution or skip to other location using GOTO.

echo Backup database
sqlcmd -S(local) -U<uid> -P<pwd> -i"backup_db.sql"
IF ERRORLEVEL 1 GOTO abort_bkp

echo Update database
sqlcmd -S(local) -U<uid> -P<pwd> -i"create_proc.sql"
IF ERRORLEVEL 1 GOTO abort

echo Update customer data
sqlcmd -S(local) -U<uid> -P<pwd> -i"update_customer_data.sql"
IF ERRORLEVEL 1 GOTO abort_with_restore

:abort_bkp
echo Error backup database. Setup aborted

:abort_with_restore
echo Error updating data. Restoring database...
sqlcmd -S(local) -U<uid> -P<pwd> -i"restore_db.sql"
IF ERRORLEVEL 1 GOTO res_falied
...
...

|||

hmm ... it seems as thought ERRORLEVEL is only set on the SUCCESS/FAILURE of the SQLCMD invocation, and not based on the SUCCESS/FAILURE of the invoked sql commands?

for example:

batch CALLBACKUP.BAT file contents:

echo Backup database
sqlcmd -SRGalbraith\SQL2005_1 -E -i"d:\backup_db.sql"
IF ERRORLEVEL 1 GOTO abort_bkp
IF ERRORLEVEL 0 GOTO done

:abort_bkp
echo Error backup database. Setup aborted

:done
echo all done now

backup_db.sql contents

backup database DataStore2 to disk = 'D:\BackupDatabase.bak'

execution results:

D:\>sqlcmd -SRGalbraith\SQL2005_1 -E -i"d:\backup_db.sql"
Msg 911, Level 16, State 11, Server RGALBRAITH\SQL2005_1, Line 1
Could not locate entry in sysdatabases for database 'DataStore2'. No entry found with that name. Make sure that the name
is entered correctly.
Msg 3013, Level 16, State 1, Server RGALBRAITH\SQL2005_1, Line 1
BACKUP DATABASE is terminating abnormally.

D:\>IF ERRORLEVEL 1 GOTO abort_bkp

D:\>IF ERRORLEVEL 0 GOTO done

D:\>echo all done now
all done now

A sample of sqlcmd failing was:

D:\>callbackup

D:\>echo Backup database
Backup database

D:\>sqlcmd -SRGalbraith\SQL2005_1 -E -i"d:\backup_db.sql"
Sqlcmd: 'd:\backup_db.sql': Invalid filename.

D:\>IF ERRORLEVEL 1 GOTO abort_bkp

D:\>echo Error backup database. Setup aborted
Error backup database. Setup aborted

D:\>echo all done now
all done now

...

As is probably obvious, I'm not much of a batch file coder :-), but the jmist of it is there - when the SQLCMD failed (file not found) then it reported error, but when the SQL script failed (database not found) no error was reported. Is there a way around that?

|||

You have to set the -b option for the SQLCMD. -b makes the batch abort with an error if the script fails. So you would write this...

@.ECHO OFF

@.echo.
@.echo Backup database
sqlcmd -S.\sqlexpress -E -i"backup_db.sql" -b
IF %ERRORLEVEL% NEQ 0 GOTO err_bkp_failed

:success
echo Database update successful
goto end

:err_bkp_failed
echo Backup failed. Aborting...
goto end

:end

HTH

|||hmm - good to know! still going to investiage the other options as well, since with the batch file I have to add a file each time.
Thanks
|||

Visual Studio .Net 2003 had a "create batch file" command which was beautiful for creating this batch file to process the sequence of sql scripts that you create.

I still use it today. But it seems we are in need to migrate to Visual Studio 2005, and this feature has been disabled now.

Do you have a more elegant solution now?

|||

Actually you don't have to modifiy the bat script each time. I have been using bat scripts to do exactly this for years.

The shell support the For Each looping structure which will set a shell variable to each file name that meet's a spec.

For Each %%1 in *.sql <execute a dos command>

I have been using the OSQL command line utility for years like this. I guess I will have to update to SQLCMD now.

You can find out the details of shell commands by going to "My Computer" <Help> and searching for "For Each"

You can find out about OSQL in BOL

|||

I've been searching solution on catching MS SQL abortion errors in a launching batch file. With option '-b', at least the batch file could return error code 1 instead of 0. Thanks for the hint!

Still, I'd appreciate if anyone could offer answer on capturing the stdout error in the batch file. My problem is that once the sql statement is aborted, it immediately exits from the erroneous line, ignores the rest code in the same script. Therefore, no error could be saved.

Also, I found that in some env. the 'sqlcmd' is not recognized (SQL Server 2000?) but 'osql' or 'isql'. Are there any differences among them (must be, but I don't know).

sql

Deploying the whole lot

Hi All,

I am adding a reporting suite to our application based upon SSRS 2005 and I am looking for a comprehensive example of how I go about deploying this.

Essentially I want to automate the entire process:

Create the DB, Create the Virtual Dirs, add my reports etc.

I have looked at the documentation and it is limited to say the least. I went down the route of attempting to script the whole thing and after 3 days of it I was still stuck on a few problems so I was hoping that someone knows of an example of how to go through all of these processes.

An example of the problems I ran into is that the "CreateVirtualDir" method on the reporting services WMI objects simply didn't work. It didn't return an error but it didn't create the Dir either!

The documentation around this method and most others is ambiguous and contains no examples.

I then went down the route of using IIS WMI to create the dirs which worked but then I was unable to update the Reporting Services configuration to pick this up without resorting to changing the config files directly - and even when I did this they didn't always seem to be picked up.

So at present instead of giving our guys in deployment a 1 click install they have a 9 page document explaining the install and set up of Reporting services from start to finish.

Any help or more thorough documentation would be greatly appreciated.

thanks,

Dhugal.

PS: Just to add I have already altered our install to install the RS module along with SQL Server so that bit is covered. I just want to script the entire configuration process.

Bump ^ ^

Dependencies are lost when alter an object

Hello,
I have an old problem but now it became ugly...
I have an object (for example a function) that is used in some views. If I am looking on dependencies, it shows me the 8 views whcih depends on it. But if I edit the function in Query Analyser and run the script (no modification on the script), it deletes
all the dependencies for the 8 views on it... The problem is worst because when I generate the initial snapshot for a merge replication, it will affect the generation order of the scripts and I get error when trying to create the views (on subscriber), s
aying that the function doesn't exist in database...
Any ideea how can I preserve the dependencies?
Thanks in advance
Catalin,
I've also wondered what is the cause of this. There was mention a while back
that it is caused by using enterprise manager to do certain changes to
objects rather than using QA. Also, it has been mentioned that renaming
objects can corrupt dependencies.
I haven't found a definitive article explaining this, but as a workaround,
if you have this problem, you might want to run sp_refreshview on your views
prior to replication.
HTH,
Paul Ibison
|||I don't work with EM, only with QA.
For example: I run this script, to create 2 functions:
CREATE FUNCTION dbo.Child () RETURNS TABLE RETURN (SELECT 1 AS Col)
GO
CREATE FUNCTION dbo._Parent () RETURNS TABLE RETURN (SELECT * FROM dbo.Child ())
GO
As you can see _Parent depens on Child.
At this moment everything is fine on dependencies (each one shows the other)
But if I execute the next script:
ALTER FUNCTION dbo.Child () RETURNS TABLE RETURN (SELECT 1 AS Col)
I loose all the referencies. And I need to generate the initial snapshot for a replication. And because no dependencies found, it will create the function in the order of names (and "_Parent" is before "Child") and when it tries to create the database (on
subscriber) I get the error because it tries to create _Parent before Child...
|||Hi Catalin,
If you don't have very many complex dependencies among functions, you may be
able to restore the dependency information by altering _Parent after you
alter Child.
HTH
-Raymond
"Catalin" <anonymous@.discussions.microsoft.com> wrote in message
news:0D8E0108-CACC-4491-A1B2-DFCD4FE641C6@.microsoft.com...
> I don't work with EM, only with QA.
> For example: I run this script, to create 2 functions:
> CREATE FUNCTION dbo.Child () RETURNS TABLE RETURN (SELECT 1 AS Col)
> GO
> CREATE FUNCTION dbo._Parent () RETURNS TABLE RETURN (SELECT * FROM
dbo.Child ())
> GO
> As you can see _Parent depens on Child.
> At this moment everything is fine on dependencies (each one shows the
other)
> But if I execute the next script:
> ALTER FUNCTION dbo.Child () RETURNS TABLE RETURN (SELECT 1 AS Col)
> I loose all the referencies. And I need to generate the initial snapshot
for a replication. And because no dependencies found, it will create the
function in the order of names (and "_Parent" is before "Child") and when it
tries to create the database (on subscriber) I get the error because it
tries to create _Parent before Child...
|||Yes, but... unfortunately I have a database with more that 1000 objects and I need to replicate it... In this case I need to "reconstruct" the dependencies tree. The simplest way that I think is to develop a small VB application that will try to re-establ
ish the order (trying each script to execute in a blank database and to determine the missing object)...

Dependancy Information

Is there a way to update the dependency information
For example if I create a SP that references a table that does not exist no record will be put in sysdepends. If I later add this table can I update the dependancies to reflect this change?the only way I can think is to run Alter procedure script
in QA, or open the SP in Enterprise Manager and add a
blank line anywhere. This will update the sysdepends table.
>--Original Message--
>Is there a way to update the dependency information?
>For example if I create a SP that references a table that
does not exist no record will be put in sysdepends. If I
later add this table can I update the dependancies to
reflect this change?
>.
>