CONVERT a text column to date format

I have a column nvarchar(8) that I need to update to a date format, MM/DD/YYYY

Some of the values have 7 characters. the rest have 8 characters as shown below:

Col1

6051998

12061999

In both rows the format is M/DD/YYYY and MM/DD/YYYY respectively. I have tried using CONVERT and CAST but receive the error:

Conversion failed when converting datetime from character string.

I've manage to generate the correct format in a select statement using CASE:

SELECT date_updted =

CASE

WHEN(selectLEN(date_updted))= 7 THEN(SELECTLEFT((RIGHT(date_updted, 7)), 1)+'/'+(SELECTLEFT((RIGHT(date_updted, 6)), 2))+'/'+(selectRIGHT(date_updted, 4)))

ELSE(SELECTLEFT((RIGHT(date_updted, 8)), 2)+'/'+(SELECTLEFT((RIGHT(date_updted, 6)), 2))+'/'+(selectRIGHT(date_updted, 4)))

END

FROM Table1

How can perform an update of this column using the UPDATE statement? I've tried the following with no success:

UPDATE dbo.Table1

SET date_updted =(SELECT date_updted =

CASE

WHEN(selectLEN(date_updted))= 7 THEN(SELECTLEFT((RIGHT(date_updted, 7)), 1)+'/'+(SELECTLEFT((RIGHT(date_updted, 6)), 2))+'/'+(selectRIGHT(date_updted, 4))As date_updted)

ELSE(SELECTLEFT((RIGHT(date_updted, 8)), 2)+'/'+(SELECTLEFT((RIGHT(date_updted, 6)), 2))+'/'+(selectRIGHT(date_updted, 4))As date_updted)

END

FROM Table1)

FROM Table1

If your data is as strongly formated as you say, you should be able to convert to datetime with something like:

Code Snippet

select aDate,
convert(datetime, right(aDate, 4) + left( right('0'+aDate, 8), 4))
as convertedDT
from ( select '6051998' as aDate union all
SELECT '12061999'
) a

/*
aDate convertedDT
--
6051998 1998-06-05 00:00:00.000
12061999 1999-12-06 00:00:00.000
*/

|||

Try something like this:

Code Snippet

DECLARE @.MyTable table
( RowID int IDENTITY,
DateCol varchar(8)
)

INSERT INTO @.MyTable VALUES ( '6051998' )
INSERT INTO @.MyTable VALUES ( '12061999' )

SELECT convert( datetime, ( stuff( stuff( right( '0' + DateCol, 8 ), 3, 0, '/' ), 6, 0, '/' )), 101 )
FROM @.MyTable

-
1998-06-05 00:00:00.000
1999-12-06 00:00:00.000

|||

Thanks for your responses and I'm sure these methods will work fine, however, my question really was how to use the UPDATE Statement to update the column in my initial post without creating a temp table or column. Maybe I missed something in your responses?

|||

Copy the expression from those select statement...

Code Snippet

UPDATE

dbo.Table1

SET

date_updted =convert(datetime, right(date_updted, 4) + left( right('0'+date_updted, 8), 4))

FROM

Table1

|||

Maybe:

Code Snippet

UPDATE dbo.Table1
SET date_updted
= convert(varchar(10),convert(datetime, right(date_updted, 4) + left( right('0'+date_updted, 8), 4)),101)

|||Very Nice! Thanks.

Conversion from char to Nchar

Hello,

I am trying to convert a single code page MS Server database into a unicode database, using the unicode data types,NCHAR, NVARCHAR, NTEXT. The problem is that in the original database, indexes and constraints have been defined on the tables whose configurations need to be changed. As a result, the ALTER TABLE command fails. Are there any other alternative solutions?
Also, data from the old database needs to be preserved. The objective is to create a unicode database which keeps the old data intact as well as accepts the new data in unicode.
It would be great if you could help!
Thanks,
Sheetal.If you have enough disk space, I strongly recommend:

1) Use SQL Enterprise Manager to script your old database
2) Edit the script to change CHAR to NCHAR
3) Edit the script to change VARCHAR to NVARCHAR
4) Edit the script to change TEXT to NTEXT
5) Create a new database
6) Play the script into the new database using SQL Query Analyzer
7) Copy the data from the old database to the new one

The down side is that this new database can take about 2.5 times as much disk space as your old database, so you have to have quite a bit of space free to make this happen.

There are other ways to do this conversion, but they are a lot more complicated. If you have the disk space, this is a much simpler way to do the conversion.

-PatP|||Hello there,

Thanks very much for your speedy reply!
Excuse me if i sound like a complete beginner, but I am not really experienced with MS Server, as a result, I'm not aware of whether my approach to scripting the database is correct or not. Is it, right click on the existing database -> All new tasks -> Generate SQL Script , and if so,then , General->Show All ;Options->All checkboxes selected??
When the script is ready, i try to execute it on the 'master' DB, after renaming the existing database.
But I get errors and it doesn't execute saying it that it doesn't recognise the user defined data types and roles (while giving their names)
How can I take care of this?
Lastly, after changing the concerned fields, i.e, char to Nchar, varchar to Nvarchar and text to Ntext, which tool is used to transfer the data from the old DB to the new empty one?

Thanks again!
Sheetal.|||That's Ok, everybody has to start somewhere!

First, create a new database. If you prefer working in a GUI environment, you can do this using SQL Enterprise Manager.

Next, open SQL Query Analyzer. Connect to the database, then open the edited script. Click on the "play" button in the toolbar, or just hit Ctrl-E to execute. The script should run in the new database with no error messages.

The simplest way to move the data is probably to use the DTS Wizard. You can get to it by right clicking the Data Transformation Service in SQL Enterprise Manager.

-PatP|||Hello Pat :-)

Many thanks again for your speedy (and warm) reply!
I tried using the SQL Enterprise Manager to create a new database, and then play into it the edited version of the old DB Script, but it gives me really wierd errors everytime, like a certain table or type doesn't exist(though it does exist in the original DB), and there's no way I can find out what's going on with the automated script generation. Any tips?
If not, I wrote a piece of code which works perfectly in converting Char to Nchar :
ALTER TABLE t_nm_reports
DROP CONSTRAINT UQ__t_nm_reports__5D60DB10

ALTER TABLE t_nm_reports
ALTER COLUMN nm_dw_name NCHAR(50) /*the column which needs to be changed*/

ALTER TABLE t_nm_reports
ADD CONSTRAINT UQ__t_nm_reports__5D60DB10
UNIQUE (nm_dw_name);

But, this is a very rough and basic way of solving the problem, done manually for each concerned table. I'm looking for a piece of code which can search for the concerned tables and perform the query, all in one program. Is that possible?
Thank you for your time!

Sheetal.|||I'm going to be really, really detailed about this. Please don't be offended, I'm trying to cover everything, not insult anyone!

1) Launch SQL Enterprise Manager
2) Navigate the tree to the server of interest
3) Double click the server to connect and open
4) Double click the Databases collection to open it
5) Right click the source database
6) Click on All Tasks | Generate SQL Script...
7) Click on the Show All button in the upper right corner
8) Click in the Script all objects checkbox
9) Click the Options tab
10) In the security section:
11) Click the checkbox for Script database users and database roles
12) Click the checkbox for Script object-level permissions
13) In the Table Scripting section:
14) Click all four checkboxes
15) Click the Ok button
16) Make the appropriate choices for saving the file

This should get you a script that includes everything, in the correct order to rebuild the schema from scratch. You should be able to edit this script to change CHAR to NCHAR and VARCHAR to NVARCHAR without any problems (or at least I can't think of any).

While you can hunt down all of the "problem child" columns and fix them as you did in your example, it is a lot more work and I'm not completely comfortable that you'll get what you really want, especially from a performance standpoint.

-PatP|||Hello there,
I'm sorry but it just doesn't seem to work :-(
Everytime I try to execute the edited script in the query analyzer(against the new empty database or the source database), I get errors like a particular table/sp doesn't exist (or incorrect syntax), even though it does exist before the execution, but seems to get dropped/deleted during the execution from the source database.
Thanks for ur time, any other suggestions would be highly appreciated!

Sheetal.|||If you have foreign key definitions (look for the keyword FOREIGN to find them), you will want to move them to the end of the script. This is due to the fact that the scripting engine doesn't always respect the "dependance sequence" of the tables, so the tables aren't always created in the same order that they were originally created.

-PatP|||You may also just be able to run your script twice, ignoring any errors that state that a particular object already exists.|||You may also just be able to run your script twice, ignoring any errors that state that a particular object already exists.As long as you skip the DROPs after the first run!

-PatP|||So, the scripting was completed successfully,with everything done as specified except that in the Options tab, the MS-DOS(OEM) format was selected, and not the Windows ANSI format.It does give some errors because of dependent objects, which when moved before the creation of the calling procedure, works fine. But it can be a hassle if they are many in number(as in my case). Any workarounds this problem??

Also, an important question for me is to know how to find and replace a certain string, eg changing char to nchar while using the query analyzer's "Replace", passes thru every string named 'varchar' or 'character' as well, so its very time consuming.I've tried using '(space)char', but its not foolproof either. Is there any way I can search for regular expressions automatically, as doin it manually in a huge database script is not very practical. Like for eg, creating a .bat file and using FINDSTR? if I'm on the right track, please guide me further!
Thanks!
Sheetal.|||Let's try a different approach...

Modify the directions for step #14 to exclude the last check box for Primary, Foreign, and Check constraints. Build the script that way (without the constraints).

Build a second script with only the constraints.

1) Launch SQL Enterprise Manager
2) Navigate the tree to the server of interest
3) Double click the server to connect and open
4) Double click the Databases collection to open it
5) Right click the source database
6) Click on All Tasks | Generate SQL Script...
7) Click on the Show All button in the upper right corner
8) Click in the Script all tables checkbox
9) Click the Formatting tab
10) Clear all of the check boxes
11) Click the Options tab
12) In the Table Scripting section:
13) Click only the PRIMARY keys, FOREIGN KEYS, and check constraints checkbox
14) Click the Ok button
15) Make the appropriate choices for saving the file

Now you should be able to play the first script, then play the second script without running into the dependancy problems you've been having.

In terms of better editing tools, I'd use an editor that recognizes regular expressions (Elvis is free, there are lots of others), or a tool like Perl that was made for those kinds of tasks.

-PatP

Conversion Error...nvarchar to Datetime

Hi Group,
I am new with SQL Server..I am working with SQL Server 2000.
I am storing the date in a nvarchar column of atable... Now I want to
show the data of Weekends..Everything is OK...But the problem is
arising with Conversion of nvarchar to date...to identify the
weekends...Like..Here DATEVALUE is a nvarchar column...But getting the
error..Value of DATEVALUE like dd-mm-yyyy...04-08-2004

------------------
Server: Msg 8115, Level 16, State 2, Line 1
Arithmetic overflow error converting expression to data type datetime.
------------------
------Actual Query----------
Select DATEVALUE,<Other Column Names> from Result where
Datepart(dw,convert(Datetime,DATEVALUE))<>1 and
Datepart(dw,convert(Datetime,DATEVALUE))<>7
------------------
Thanks in advance..
Regards
Arijit Chatterjee(arijitchatterjee123@.yahoo.co.in) writes:
> I am new with SQL Server..I am working with SQL Server 2000.
> I am storing the date in a nvarchar column of atable... Now I want to
> show the data of Weekends..Everything is OK...But the problem is
> arising with Conversion of nvarchar to date...to identify the
> weekends...Like..Here DATEVALUE is a nvarchar column...But getting the
> error..Value of DATEVALUE like dd-mm-yyyy...04-08-2004

Best is to store date values in datetime columns. If you use character
format, you should use char (the n just doubles the space with no gain
for it, and the var is pointless since size is fixed), and you should use
the format YYYYMMDD. Furthermore, you should attach a constraint to the
columns

datecol char(8) CONSTRAINT ck_tbl_datecol CHECK (isdate(datecol) = 1)

to ascertain that you don't get illegal values.

Storing dates in a format like DD-MM-YYYY is going to give all sorts of
headache. 04-08-2004 could be interpreted as Aug 4th or April 8th, depending
on language and datefromat settings. (And, in case of humans, of the
perceptions of the user.) You can't sort on this format (unless you really
want 3 Aug to come before 4 June).

The format YYYYMMDD sorts well, and is always interpreted in the same way.

See also http://www.karaszi.com/SQLServer/info_datetime.asp.

--
Erland Sommarskog, SQL Server MVP, esquel@.sommarskog.se

Books Online for SQL Server SP3 at
http://www.microsoft.com/sql/techin.../2000/books.asp|||See this

declare @.t table (d varchar(20))
insert into @.t values('10-apr-2005')
insert into @.t values('10-MAy-2005')
insert into @.t values('10-Jun-2005')
insert into @.t values('10-Jul-2005')
Select * from @.t where Datepart(dw,convert(Datetime,d))<>1 and
Datepart(dw,convert(Datetime,d))<>7

Madhivanan|||Thanks for your great help..
Regards
Arijit Chatterjee

Conversion error when calling stored procedure

Hi all,
I want to execute a stored procedure from a report. The stored procedure
takes nvarchar(50) parameters. When I call the procedure from a report with
string paramerers, I get the error 'Implicit conversion from data type
sql_variant to varchar is not allowed'. But, I cannot Cast or Convert my
reporting services parameters when calling Exec to run the stored procedure.
Help! Anyone experience this kind of problem before? Any suggestions are
welcome!hey.
Not sure why it happens as I get those all the time to. I have found that
if i create a make the first dataset something simple, like the query to
populate a parameter drop down list, then add a second data source I can then
call the execute statement for the proc. Boqus I know but it works. Buggy
software is my quess, I think it may have been the service pack as I did not
do this a few months back.
hth
"Bas" wrote:
> Hi all,
> I want to execute a stored procedure from a report. The stored procedure
> takes nvarchar(50) parameters. When I call the procedure from a report with
> string paramerers, I get the error 'Implicit conversion from data type
> sql_variant to varchar is not allowed'. But, I cannot Cast or Convert my
> reporting services parameters when calling Exec to run the stored procedure.
> Help! Anyone experience this kind of problem before? Any suggestions are
> welcome!
>

Conversion

Hi,
Can I convert an nvarchar field to numeric type?
Thanks
Chandrayes
thanks,
Jose de Jesus Jr. Mcp,Mcdba
Data Architect
Sykes Asia (Manila philippines)
MCP #2324787
"Chandra" wrote:

> Hi,
> Can I convert an nvarchar field to numeric type?
> Thanks
> Chandra|||try
Cast (myNvarCharTableColumn as numeric)
"Chandra" wrote:

> Hi,
> Can I convert an nvarchar field to numeric type?
> Thanks
> Chandra|||Hi,
declare @.t nvarchar(10)
set @.t='10'
select convert(numeric(5,2),@.t)
Ensure that you have no character data in te column
Thanks
Hari
SQL Server MVP
"RioDD" <RioDD@.discussions.microsoft.com> wrote in message
news:641160B8-BA5D-4FF4-92E3-76C39B2FA79B@.microsoft.com...
> try
> Cast (myNvarCharTableColumn as numeric)
> "Chandra" wrote:
>|||Also look at this recent thread for some details on numeric data types:
http://msdn.microsoft.com/newsgroup...71-b7768b43fcf3
...in case you missed it.
ML|||I tried Hari's solution but it is giving the following error.
Server: Msg 8114, Level 16, State 5, Line 1
Error converting data type nvarchar to numeric.
what could be the reason?
Thanks
Chandra
"Hari Pra" wrote:

> Hi,
> declare @.t nvarchar(10)
> set @.t='10'
> select convert(numeric(5,2),@.t)
> Ensure that you have no character data in te column
> Thanks
> Hari
> SQL Server MVP
> "RioDD" <RioDD@.discussions.microsoft.com> wrote in message
> news:641160B8-BA5D-4FF4-92E3-76C39B2FA79B@.microsoft.com...
>
>|||Obviously your nvarchar column contains values that cannot be converted to
the numeric data type. Read my other post in this thread.
It's imperative that you test values before attempting to cast/convert them,
since in some circumstances conversions even if successful can lead to
unexpected results.
And when it comes to numbers certain aspects of life can become devastated
by such faults.
ML|||>I tried Hari's solution but it is giving the following error.
> Server: Msg 8114, Level 16, State 5, Line 1
> Error converting data type nvarchar to numeric.
> what could be the reason?
The reason is that you have non-numeric data in your NVARCHAR column. How
is this supposed to work?
SELECT CONVERT(NUMERIC(5,2), 'fooblat$%xpq@.g')
?|||Don't you know that 'fooblat$%xpq@.g' = 42?
:)
ML|||Of course, I do. But I haven't applied that specific hotfix to my
production SQL Servers yet. :-)
"ML" <ML@.discussions.microsoft.com> wrote in message
news:D5B6DD32-14D3-4413-998F-8F2789EE7747@.microsoft.com...
> Don't you know that 'fooblat$%xpq@.g' = 42?
> :)
>
> ML

Conversion

Hi,
I am in the process writing scripts for CHAR/VARCHAR to
NCHAR/NVARCHAR conversion..Infact i have completed
it..When I run these scripts against DB is verly
slow..This is happening when i run convert scripts
eaxctly..Even select name from sysobjects doesn't return
values..There are heavy IO is going on the DB..what are
the areas i should concentrate the to improve performance..
Basically I have written cursor whic executes one by one
the alter scripts..
Sridhar...Changing a CHAR to VARCHAR will most likely involve each row being affected
and hence all the I/O. If you are altering more than one column per table
you might get better performance by creating a second table with the changed
columns and Inserting all the rows into it. Drop the original, rename the
new to the old and add the appropriate indexes, RI etc.. Just make sure you
have good backups first.
Andrew J. Kelly SQL MVP
<anonymous@.discussions.microsoft.com> wrote in message
news:297601c47e07$62bd4ae0$a501280a@.phx.gbl...
> Hi,
> I am in the process writing scripts for CHAR/VARCHAR to
> NCHAR/NVARCHAR conversion..Infact i have completed
> it..When I run these scripts against DB is verly
> slow..This is happening when i run convert scripts
> eaxctly..Even select name from sysobjects doesn't return
> values..There are heavy IO is going on the DB..what are
> the areas i should concentrate the to improve performance..
>
> Basically I have written cursor whic executes one by one
> the alter scripts..
>
> Sridhar...
>

Conversion

Hi,
I am in the process writing scripts for CHAR/VARCHAR to
NCHAR/NVARCHAR conversion..Infact i have completed
it..When I run these scripts against DB is verly
slow..This is happening when i run convert scripts
eaxctly..Even select name from sysobjects doesn't return
values..There are heavy IO is going on the DB..what are
the areas i should concentrate the to improve performance..
Basically I have written cursor whic executes one by one
the alter scripts..
Sridhar...
Changing a CHAR to VARCHAR will most likely involve each row being affected
and hence all the I/O. If you are altering more than one column per table
you might get better performance by creating a second table with the changed
columns and Inserting all the rows into it. Drop the original, rename the
new to the old and add the appropriate indexes, RI etc.. Just make sure you
have good backups first.
Andrew J. Kelly SQL MVP
<anonymous@.discussions.microsoft.com> wrote in message
news:297601c47e07$62bd4ae0$a501280a@.phx.gbl...
> Hi,
> I am in the process writing scripts for CHAR/VARCHAR to
> NCHAR/NVARCHAR conversion..Infact i have completed
> it..When I run these scripts against DB is verly
> slow..This is happening when i run convert scripts
> eaxctly..Even select name from sysobjects doesn't return
> values..There are heavy IO is going on the DB..what are
> the areas i should concentrate the to improve performance..
>
> Basically I have written cursor whic executes one by one
> the alter scripts..
>
> Sridhar...
>

Conversion

Hi,
I am in the process writing scripts for CHAR/VARCHAR to
NCHAR/NVARCHAR conversion..Infact i have completed
it..When I run these scripts against DB is verly
slow..This is happening when i run convert scripts
eaxctly..Even select name from sysobjects doesn't return
values..There are heavy IO is going on the DB..what are
the areas i should concentrate the to improve performance..
Basically I have written cursor whic executes one by one
the alter scripts..
Sridhar...Can you tell us how you are performing the upgrade i.e.
Have a different database or are you performing an alter
column ?
Thanks
Peter
>--Original Message--
>Hi,
> I am in the process writing scripts for CHAR/VARCHAR
to
>NCHAR/NVARCHAR conversion..Infact i have completed
>it..When I run these scripts against DB is verly
>slow..This is happening when i run convert scripts
>eaxctly..Even select name from sysobjects doesn't return
>values..There are heavy IO is going on the DB..what are
>the areas i should concentrate the to improve
performance..
>
>Basically I have written cursor whic executes one by one
>the alter scripts..
>
>Sridhar...
>.
>|||By alter columns...
>--Original Message--
>Can you tell us how you are performing the upgrade i.e.
>Have a different database or are you performing an alter
>column ?
>Thanks
>Peter
>>--Original Message--
>>Hi,
>> I am in the process writing scripts for CHAR/VARCHAR
>to
>>NCHAR/NVARCHAR conversion..Infact i have completed
>>it..When I run these scripts against DB is verly
>>slow..This is happening when i run convert scripts
>>eaxctly..Even select name from sysobjects doesn't return
>>values..There are heavy IO is going on the DB..what are
>>the areas i should concentrate the to improve
>performance..
>>
>>Basically I have written cursor whic executes one by one
>>the alter scripts..
>>
>>Sridhar...
>>.
>.
>|||Changing a CHAR to VARCHAR will most likely involve each row being affected
and hence all the I/O. If you are altering more than one column per table
you might get better performance by creating a second table with the changed
columns and Inserting all the rows into it. Drop the original, rename the
new to the old and add the appropriate indexes, RI etc.. Just make sure you
have good backups first.
--
Andrew J. Kelly SQL MVP
<anonymous@.discussions.microsoft.com> wrote in message
news:297601c47e07$62bd4ae0$a501280a@.phx.gbl...
> Hi,
> I am in the process writing scripts for CHAR/VARCHAR to
> NCHAR/NVARCHAR conversion..Infact i have completed
> it..When I run these scripts against DB is verly
> slow..This is happening when i run convert scripts
> eaxctly..Even select name from sysobjects doesn't return
> values..There are heavy IO is going on the DB..what are
> the areas i should concentrate the to improve performance..
>
> Basically I have written cursor whic executes one by one
> the alter scripts..
>
> Sridhar...
>|||Consider what Andrew has said, but also regardless of
which way you are doing it take off all indexes before you
start.
How many rows of data are you converting ?
>--Original Message--
>By alter columns...
>>--Original Message--
>>Can you tell us how you are performing the upgrade i.e.
>>Have a different database or are you performing an alter
>>column ?
>>Thanks
>>Peter
>>--Original Message--
>>Hi,
>> I am in the process writing scripts for CHAR/VARCHAR
>>to
>>NCHAR/NVARCHAR conversion..Infact i have completed
>>it..When I run these scripts against DB is verly
>>slow..This is happening when i run convert scripts
>>eaxctly..Even select name from sysobjects doesn't
return
>>values..There are heavy IO is going on the DB..what are
>>the areas i should concentrate the to improve
>>performance..
>>
>>Basically I have written cursor whic executes one by
one
>>the alter scripts..
>>
>>Sridhar...
>>.
>>.
>.
>

CONTAINS

Hi,
I have a table like this:
CREATE TABLE T1(
C1 int identity(1,1) PRIMARY KEY,
C2 NVARCHAR(50))
INSERT T1(C2) VALUES('test')
INSERT T1(C2) VALUES('Xtest')
INSERT T1(C2) VALUES('testX')
Assuming that C2 is enabled for FTS, this query does not return "Xtest":
SELECT * FROM T1 WHERE CONTAINS (*,'"*test*"')
It returns only "test" and "testX". How can I do that?
Thanks in advance,
Leila
"Leila" <Leilas@.hotpop.com> wrote in message
news:O7ujWH2iFHA.3300@.TK2MSFTNGP10.phx.gbl...
> Hi,
> I have a table like this:
> CREATE TABLE T1(
> C1 int identity(1,1) PRIMARY KEY,
> C2 NVARCHAR(50))
> INSERT T1(C2) VALUES('test')
> INSERT T1(C2) VALUES('Xtest')
> INSERT T1(C2) VALUES('testX')
> Assuming that C2 is enabled for FTS, this query does not return "Xtest":
> SELECT * FROM T1 WHERE CONTAINS (*,'"*test*"')
> It returns only "test" and "testX". How can I do that?
You can't, unless you use "Xtest" or "Xtest*". FTS doesn't handle suffix
searches, and so the leading * is ignored. Another option would be to fall
back to LIKE when a leading * is used.
Dan
|||I think you mean prefix (comes before) which SQL FTS does not support. SQL
FTS does support suffix (comes at the end) type searches when you use the
wildcard operator in the Contains predicate.
Hilary Cotter
Looking for a SQL Server replication book?
http://www.nwsu.com/0974973602.html
Looking for a FAQ on Indexing Services/SQL FTS
http://www.indexserverfaq.com
"Daniel Crichton" <msnews@.worldofspack.co.uk> wrote in message
news:e1DACM3iFHA.3216@.TK2MSFTNGP10.phx.gbl...
> "Leila" <Leilas@.hotpop.com> wrote in message
> news:O7ujWH2iFHA.3300@.TK2MSFTNGP10.phx.gbl...
> You can't, unless you use "Xtest" or "Xtest*". FTS doesn't handle suffix
> searches, and so the leading * is ignored. Another option would be to fall
> back to LIKE when a leading * is used.
> Dan
>
|||"Hilary Cotter" <hilary.cotter@.gmail.com> wrote in message
news:%23RPqSH4iFHA.3064@.TK2MSFTNGP15.phx.gbl...
>I think you mean prefix (comes before) which SQL FTS does not support. SQL
> FTS does support suffix (comes at the end) type searches when you use the
> wildcard operator in the Contains predicate.
Possibly, I can't remember which way around they go. If it refers to the
string part, it's suffix that isn't supported as the wildcard will be the
prefix. Thanks for pointing out my error though
Dan
|||Thank you all :-)
Then LIKE will be the only poosible way?
"Daniel Crichton" <msnews@.worldofspack.co.uk> wrote in message
news:ek0j$34iFHA.3216@.TK2MSFTNGP10.phx.gbl...[vbcol=seagreen]
> "Hilary Cotter" <hilary.cotter@.gmail.com> wrote in message
> news:%23RPqSH4iFHA.3064@.TK2MSFTNGP15.phx.gbl...
SQL[vbcol=seagreen]
the
> Possibly, I can't remember which way around they go. If it refers to the
> string part, it's suffix that isn't supported as the wildcard will be the
> prefix. Thanks for pointing out my error though
> Dan
>
|||Unfortunately so.
Hilary Cotter
Looking for a SQL Server replication book?
http://www.nwsu.com/0974973602.html
Looking for a FAQ on Indexing Services/SQL FTS
http://www.indexserverfaq.com
"Leila" <Leilas@.hotpop.com> wrote in message
news:ehN0lR%23iFHA.576@.tk2msftngp13.phx.gbl...[vbcol=seagreen]
> Thank you all :-)
> Then LIKE will be the only poosible way?
>
> "Daniel Crichton" <msnews@.worldofspack.co.uk> wrote in message
> news:ek0j$34iFHA.3216@.TK2MSFTNGP10.phx.gbl...
> SQL
> the
the
>

constraints and altering tables

I'm in the process of trying to convert a database such that all the strings
(VARCHAR) are converted to wide strings (NVARCHAR). I have a script that
accomplishes this by removing all the primary key constraints, converts the
necessary columns, and then replaces the constraints. The script walks the
sysnames table and stores all the constraints in a table variable, and
constructs a script to recreate all the constraings based on the 'xtype'
column from sysindexes (this is based on the system stored procedure
sp_pkeys). The script creates a constraint if the xtype is of type 'PK', or
creates an index based on the INDEXPROPERTY of the index, whether it be
unique, and either clustered or non-clustered.
This works for the most part, but I have found that there are constraints
being created on some columns that did not exist before the conversion. For
example, I have a table which has a primary key on it's identity columns
defined to automatically insert a new value at each insert incremented by 1.
After the conversion, there is an additional constraint placed on this table
which prevents a value of NULL from being added, which should be a problem
due to the IDENTITY column, yet attempting to do an insert on this table
generates an error saying that a NULL value cannot be inserted. I'm not
manually inserting anything, this should just bump the id value by one and
do the insert, but this new constraint prevents this, leaving me with a
table I can no longer insert into.
In another case, I have several varchar columns that have default
constraints (simple text strings), which are also dropped before conversion.
Upon replacing the constraints read from sysnames, I get similar errors
regarding not being able to insert nulls into these columns, which I didn't
get before, as these columns had default values.
My questions are, is it possible to exactly recreate constraints
programmatically? Is there a preffered method for converting databases from
narrow to wide character?
Thanks for any advice,
-Gary> This works for the most part, but I have found that there are constraints
> being created on some columns that did not exist before the conversion.
> For
> example, I have a table which has a primary key on it's identity columns
> defined to automatically insert a new value at each insert incremented by
> 1.
> After the conversion, there is an additional constraint placed on this
> table
> which prevents a value of NULL from being added, which should be a problem
> due to the IDENTITY column, yet attempting to do an insert on this table
> generates an error saying that a NULL value cannot be inserted. I'm not
> manually inserting anything, this should just bump the id value by one and
> do the insert, but this new constraint prevents this, leaving me with a
> table I can no longer insert into.
Is the IDENTITY property still set for the column? If so, then what does
your INSERT statement look like? If not, then your script is, well,
imperfect.|||Yes, It is still set. the insert statement does inserts using enumerated
column names.
INSERT INTO RptTable (
ShortName, FullName, Description,
StockReport, ReportFilename,
AutoprintReport, PrintDaily, PrintWly, PrintMonthly,
ExportReport, ExportDaily, ExportWly, ExportMonthly, ExportFormat,
DisplayOrder, MinDateVariable, MaxDateVariable, Visible,
ReportCategoryId, HelpFilename, CLCompliant,
Param1_Label, Param1_Description)
VALUES
(@.InstallReport_ShortName, @.InstallReport_FullName,
@.InstallReport_Description,
0, @.InstallReport_ReportFilename,
0, 0, 0, 0,
0, 0, 0, 0, @.InstallReport_ExportFormat,
50, @.InstallReport_MinDateVariable, @.InstallReport_MaxDateVariable,
@.InstallReport_Visible,
dbo.idw_get_report_category_id( @.InstallCategory_ShortName ),
@.InstallReport_HelpFilename, 0,
@.InstallReport_Var1Title, @.InstallReport_Var1Description
);
The identity is on a column called ReportId, which is defined as (at table
creation time)
ReportId int PRIMARY KEY NOT NULL IDENTITY(1,1)
Thanks.
-Gary
"Aaron Bertrand [SQL Server MVP]" <ten.xoc@.dnartreb.noraa> wrote in message
news:O39thSqzFHA.1256@.TK2MSFTNGP09.phx.gbl...
> Is the IDENTITY property still set for the column? If so, then what does
> your INSERT statement look like? If not, then your script is, well,
> imperfect.
>
>|||Gary,
Perhaps there is an INSTEAD OF trigger on the table (which requires that
you specify values for all non-null columns that have no DEFAULT constraint,
even if those values will not ultimately be inserted anywhere), or you have
IDENTITY_INSERT set to ON for this table.
If these guesses are wrong, please cut and paste the error message you get.
Steve Kass
Drew University
Gary wrote:

>Yes, It is still set. the insert statement does inserts using enumerated
>column names.
>INSERT INTO RptTable (
> ShortName, FullName, Description,
> StockReport, ReportFilename,
> AutoprintReport, PrintDaily, PrintWly, PrintMonthly,
> ExportReport, ExportDaily, ExportWly, ExportMonthly, ExportFormat,
> DisplayOrder, MinDateVariable, MaxDateVariable, Visible,
> ReportCategoryId, HelpFilename, CLCompliant,
> Param1_Label, Param1_Description)
>VALUES
> (@.InstallReport_ShortName, @.InstallReport_FullName,
>@.InstallReport_Description,
> 0, @.InstallReport_ReportFilename,
> 0, 0, 0, 0,
> 0, 0, 0, 0, @.InstallReport_ExportFormat,
> 50, @.InstallReport_MinDateVariable, @.InstallReport_MaxDateVariable,
>@.InstallReport_Visible,
> dbo.idw_get_report_category_id( @.InstallCategory_ShortName ),
>@.InstallReport_HelpFilename, 0,
> @.InstallReport_Var1Title, @.InstallReport_Var1Description
> );
>The identity is on a column called ReportId, which is defined as (at table
>creation time)
>ReportId int PRIMARY KEY NOT NULL IDENTITY(1,1)
>Thanks.
>-Gary
>"Aaron Bertrand [SQL Server MVP]" <ten.xoc@.dnartreb.noraa> wrote in message
>news:O39thSqzFHA.1256@.TK2MSFTNGP09.phx.gbl...
>
>
>