I need help with next problem:
Exist table my_rows with next columns:
id - int identity
name - varchar
is_locked - bit
I need create store procedure which will return every time diferent
row for every request from different processes.
the pseudo code for the procedure :
1 select statement is :
select @.id = selct top 1 id from my_rows where is_locked =0
2 update my_rows set is_locked = 1 where id=@.id
3 return @.id
In different words : if in same time I call this procedure from 2
different connections, it will return 2 different record.
I also want that first call to procedure will not generate lock error
for second call in same time(just second call will wait for finish
first , it is ok.)
ThanksHi
You select and update statement should be contained within one transaction
and you can use the UPDLOCK hint on the select statement to stop others
returning that value
CREATE PROCEDURE GetLock ( @.id int OUTPUT ) AS
SET NOCOUNT ON
BEGIN TRANSACTION
SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) WHERE is_locked =0 )
-- Error checking
UPDATE my_rows SET is_locked = 1 WHERE id=@.id
-- Error checking
COMMIT TRANSACTION
RETURN
Alternatively you could use
UPDATE my_rows
SET is_locked = 1
WHERE id = (SELECT MAX(id) FROM my_rows WHERE is_locked = 0)
but you would not know the ID that was updated.
Using this type of locking can potentially cause a bottleneck and poor
performance.
John
"is_vlb50@.hotmail.com" wrote:
> I need help with next problem:
> Exist table my_rows with next columns:
> id - int identity
> name - varchar
> is_locked - bit
> I need create store procedure which will return every time diferent
> row for every request from different processes.
> the pseudo code for the procedure :
> 1 select statement is :
> select @.id = selct top 1 id from my_rows where is_locked =0
> 2 update my_rows set is_locked = 1 where id=@.id
> 3 return @.id
> In different words : if in same time I call this procedure from 2
> different connections, it will return 2 different record.
> I also want that first call to procedure will not generate lock error
> for second call in same time(just second call will wait for finish
> first , it is ok.)
> Thanks
>|||On Sep 26, 10:12 am, John Bell <jbellnewspo...@.hotmail.com> wrote:
> Hi
> You select and update statement should be contained within one transaction
> and you can use the UPDLOCK hint on the select statement to stop others
> returning that value
> CREATE PROCEDURE GetLock ( @.id int OUTPUT ) AS
> SET NOCOUNT ON
> BEGIN TRANSACTION
> SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) WHERE is_locked =0 )
> -- Error checking
> UPDATE my_rows SET is_locked = 1 WHERE id=@.id
> -- Error checking
> COMMIT TRANSACTION
> RETURN
> Alternatively you could use
> UPDATE my_rows
> SET is_locked = 1
> WHERE id = (SELECT MAX(id) FROM my_rows WHERE is_locked = 0)
> but you would not know the ID that was updated.
> Using this type of locking can potentially cause a bottleneck and poor
> performance.
> John
>
> "is_vl...@.hotmail.com" wrote:
> > I need help with next problem:
> > Exist table my_rows with next columns:
> > id - int identity
> > name - varchar
> > is_locked - bit
> > I need create store procedure which will return every time diferent
> > row for every request from different processes.
> > the pseudo code for the procedure :
> > 1 select statement is :
> > select @.id = selct top 1 id from my_rows where is_locked =0
> > 2 update my_rows set is_locked = 1 where id=@.id
> > 3 return @.id
> > In different words : if in same time I call this procedure from 2
> > different connections, it will return 2 different record.
> > I also want that first call to procedure will not generate lock error
> > for second call in same time(just second call will wait for finish
> > first , it is ok.)
> > Thanks- Hide quoted text -
> - Show quoted text -
You wroute:
SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) WHERE is_locked
=0 )
-- Error checking
Could you explain what is reason for "Error checking " and what I can
do.
If some body in same time call same procedure , I can get error ?
Thanks|||Hi
Read http://www.sommarskog.se/error-handling-I.html and
http://www.sommarskog.se/error-handling-II.html which gives you a very good
idea of what to do. If you follow Erlands advice you would also have checked
for errors when you commit the transaction and have code to rollback.
John
"is_vlb50@.hotmail.com" wrote:
> On Sep 26, 10:12 am, John Bell <jbellnewspo...@.hotmail.com> wrote:
> > Hi
> >
> > You select and update statement should be contained within one transaction
> > and you can use the UPDLOCK hint on the select statement to stop others
> > returning that value
> >
> > CREATE PROCEDURE GetLock ( @.id int OUTPUT ) AS
> > SET NOCOUNT ON
> > BEGIN TRANSACTION
> > SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) WHERE is_locked =0 )
> > -- Error checking
> > UPDATE my_rows SET is_locked = 1 WHERE id=@.id
> > -- Error checking
> > COMMIT TRANSACTION
> > RETURN
> >
> > Alternatively you could use
> >
> > UPDATE my_rows
> > SET is_locked = 1
> > WHERE id = (SELECT MAX(id) FROM my_rows WHERE is_locked = 0)
> >
> > but you would not know the ID that was updated.
> >
> > Using this type of locking can potentially cause a bottleneck and poor
> > performance.
> >
> > John
> >
> >
> >
> > "is_vl...@.hotmail.com" wrote:
> > > I need help with next problem:
> > > Exist table my_rows with next columns:
> > > id - int identity
> > > name - varchar
> > > is_locked - bit
> >
> > > I need create store procedure which will return every time diferent
> > > row for every request from different processes.
> >
> > > the pseudo code for the procedure :
> > > 1 select statement is :
> > > select @.id = selct top 1 id from my_rows where is_locked =0
> > > 2 update my_rows set is_locked = 1 where id=@.id
> > > 3 return @.id
> >
> > > In different words : if in same time I call this procedure from 2
> > > different connections, it will return 2 different record.
> > > I also want that first call to procedure will not generate lock error
> > > for second call in same time(just second call will wait for finish
> > > first , it is ok.)
> > > Thanks- Hide quoted text -
> >
> > - Show quoted text -
> You wroute:
> SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) WHERE is_locked
> =0 )
> -- Error checking
> Could you explain what is reason for "Error checking " and what I can
> do.
> If some body in same time call same procedure , I can get error ?
> Thanks
>
>
>|||On Sep 26, 6:02 pm, John Bell <jbellnewspo...@.hotmail.com> wrote:
> Hi
> Readhttp://www.sommarskog.se/error-handling-I.htmlandhttp://www.sommarskog.se/error-handling-II.htmlwhich gives you a very good
> idea of what to do. If you follow Erlands advice you would also have checked
> for errors when you commit the transaction and have code to rollback.
> John
>
> "is_vl...@.hotmail.com" wrote:
> > On Sep 26, 10:12 am, John Bell <jbellnewspo...@.hotmail.com> wrote:
> > > Hi
> > > You select and update statement should be contained within one transaction
> > > and you can use the UPDLOCK hint on the select statement to stop others
> > > returning that value
> > > CREATE PROCEDURE GetLock ( @.id int OUTPUT ) AS
> > > SET NOCOUNT ON
> > > BEGIN TRANSACTION
> > > SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) WHERE is_locked =0 )
> > > -- Error checking
> > > UPDATE my_rows SET is_locked = 1 WHERE id=@.id
> > > -- Error checking
> > > COMMIT TRANSACTION
> > > RETURN
> > > Alternatively you could use
> > > UPDATE my_rows
> > > SET is_locked = 1
> > > WHERE id = (SELECT MAX(id) FROM my_rows WHERE is_locked = 0)
> > > but you would not know the ID that was updated.
> > > Using this type of locking can potentially cause a bottleneck and poor
> > > performance.
> > > John
> > > "is_vl...@.hotmail.com" wrote:
> > > > I need help with next problem:
> > > > Exist table my_rows with next columns:
> > > > id - int identity
> > > > name - varchar
> > > > is_locked - bit
> > > > I need create store procedure which will return every time diferent
> > > > row for every request from different processes.
> > > > the pseudo code for the procedure :
> > > > 1 select statement is :
> > > > select @.id = selct top 1 id from my_rows where is_locked =0
> > > > 2 update my_rows set is_locked = 1 where id=@.id
> > > > 3 return @.id
> > > > In different words : if in same time I call this procedure from 2
> > > > different connections, it will return 2 different record.
> > > > I also want that first call to procedure will not generate lock error
> > > > for second call in same time(just second call will wait for finish
> > > > first , it is ok.)
> > > > Thanks- Hide quoted text -
> > > - Show quoted text -
> > You wroute:
> > SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) WHERE is_locked
> > =0 )
> > -- Error checking
> > Could you explain what is reason for "Error checking " and what I can
> > do.
> > If some body in same time call same procedure , I can get error ?
> > Thanks- Hide quoted text -
> - Show quoted text -
Thanks,
it was very helpful. One last question:
if statetement "SELECT top 1 id FROM my_rows (UPDLOCK) ..." in second
call in store procedure can raise error because first call in first SP
still running (still locks the record) or it just will receive another
row?
Thanks|||> if statetement "SELECT top 1 id FROM my_rows (UPDLOCK) ..." in second
> call in store procedure can raise error because first call in first SP
> still running (still locks the record) or it just will receive another
> row?
TOP 1 can give you *any* row. If the row decided for is locked already by an incompatible lock, then
you will be blocked. If you want some other row, you might wan to check out the READPAST hint.
--
Tibor Karaszi, SQL Server MVP
http://www.karaszi.com/sqlserver/default.asp
http://sqlblog.com/blogs/tibor_karaszi
<is_vlb50@.hotmail.com> wrote in message news:1190831328.803233.199140@.19g2000hsx.googlegroups.com...
> On Sep 26, 6:02 pm, John Bell <jbellnewspo...@.hotmail.com> wrote:
>> Hi
>> Readhttp://www.sommarskog.se/error-handling-I.htmlandhttp://www.sommarskog.se/error-handling-II.htmlwhich
>> gives you a very good
>> idea of what to do. If you follow Erlands advice you would also have checked
>> for errors when you commit the transaction and have code to rollback.
>> John
>>
>> "is_vl...@.hotmail.com" wrote:
>> > On Sep 26, 10:12 am, John Bell <jbellnewspo...@.hotmail.com> wrote:
>> > > Hi
>> > > You select and update statement should be contained within one transaction
>> > > and you can use the UPDLOCK hint on the select statement to stop others
>> > > returning that value
>> > > CREATE PROCEDURE GetLock ( @.id int OUTPUT ) AS
>> > > SET NOCOUNT ON
>> > > BEGIN TRANSACTION
>> > > SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) WHERE is_locked =0 )
>> > > -- Error checking
>> > > UPDATE my_rows SET is_locked = 1 WHERE id=@.id
>> > > -- Error checking
>> > > COMMIT TRANSACTION
>> > > RETURN
>> > > Alternatively you could use
>> > > UPDATE my_rows
>> > > SET is_locked = 1
>> > > WHERE id = (SELECT MAX(id) FROM my_rows WHERE is_locked = 0)
>> > > but you would not know the ID that was updated.
>> > > Using this type of locking can potentially cause a bottleneck and poor
>> > > performance.
>> > > John
>> > > "is_vl...@.hotmail.com" wrote:
>> > > > I need help with next problem:
>> > > > Exist table my_rows with next columns:
>> > > > id - int identity
>> > > > name - varchar
>> > > > is_locked - bit
>> > > > I need create store procedure which will return every time diferent
>> > > > row for every request from different processes.
>> > > > the pseudo code for the procedure :
>> > > > 1 select statement is :
>> > > > select @.id = selct top 1 id from my_rows where is_locked =0
>> > > > 2 update my_rows set is_locked = 1 where id=@.id
>> > > > 3 return @.id
>> > > > In different words : if in same time I call this procedure from 2
>> > > > different connections, it will return 2 different record.
>> > > > I also want that first call to procedure will not generate lock error
>> > > > for second call in same time(just second call will wait for finish
>> > > > first , it is ok.)
>> > > > Thanks- Hide quoted text -
>> > > - Show quoted text -
>> > You wroute:
>> > SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) WHERE is_locked
>> > =0 )
>> > -- Error checking
>> > Could you explain what is reason for "Error checking " and what I can
>> > do.
>> > If some body in same time call same procedure , I can get error ?
>> > Thanks- Hide quoted text -
>> - Show quoted text -
> Thanks,
> it was very helpful. One last question:
> if statetement "SELECT top 1 id FROM my_rows (UPDLOCK) ..." in second
> call in store procedure can raise error because first call in first SP
> still running (still locks the record) or it just will receive another
> row?
> Thanks
>|||On Sep 26, 8:40 pm, "Tibor Karaszi"
<tibor_please.no.email_kara...@.hotmail.nomail.com> wrote:
> > if statetement "SELECT top 1 id FROM my_rows (UPDLOCK) ..." in second
> > call in store procedure can raise error because first call in first SP
> > still running (still locks the record) or it just will receive another
> > row?
> TOP 1 can give you *any* row. If the row decided for is locked already by an incompatible lock, then
> you will be blocked. If you want some other row, you might wan to check out the READPAST hint.
> --
> Tibor Karaszi, SQL Server MVPhttp://www.karaszi.com/sqlserver/default.asphttp://sqlblog.com/blogs/tibor_karaszi
>
> <is_vl...@.hotmail.com> wrote in messagenews:1190831328.803233.199140@.19g2000hsx.googlegroups.com...
> > On Sep 26, 6:02 pm, John Bell <jbellnewspo...@.hotmail.com> wrote:
> >> Hi
> >> Readhttp://www.sommarskog.se/error-handling-I.htmlandhttp://www.sommarsko...
> >> gives you a very good
> >> idea of what to do. If you follow Erlands advice you would also have checked
> >> for errors when you commit the transaction and have code to rollback.
> >> John
> >> "is_vl...@.hotmail.com" wrote:
> >> > On Sep 26, 10:12 am, John Bell <jbellnewspo...@.hotmail.com> wrote:
> >> > > Hi
> >> > > You select and update statement should be contained within one transaction
> >> > > and you can use the UPDLOCK hint on the select statement to stop others
> >> > > returning that value
> >> > > CREATE PROCEDURE GetLock ( @.id int OUTPUT ) AS
> >> > > SET NOCOUNT ON
> >> > > BEGIN TRANSACTION
> >> > > SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) WHERE is_locked =0 )
> >> > > -- Error checking
> >> > > UPDATE my_rows SET is_locked = 1 WHERE id=@.id
> >> > > -- Error checking
> >> > > COMMIT TRANSACTION
> >> > > RETURN
> >> > > Alternatively you could use
> >> > > UPDATE my_rows
> >> > > SET is_locked = 1
> >> > > WHERE id = (SELECT MAX(id) FROM my_rows WHERE is_locked = 0)
> >> > > but you would not know the ID that was updated.
> >> > > Using this type of locking can potentially cause a bottleneck and poor
> >> > > performance.
> >> > > John
> >> > > "is_vl...@.hotmail.com" wrote:
> >> > > > I need help with next problem:
> >> > > > Exist table my_rows with next columns:
> >> > > > id - int identity
> >> > > > name - varchar
> >> > > > is_locked - bit
> >> > > > I need create store procedure which will return every time diferent
> >> > > > row for every request from different processes.
> >> > > > the pseudo code for the procedure :
> >> > > > 1 select statement is :
> >> > > > select @.id = selct top 1 id from my_rows where is_locked =0
> >> > > > 2 update my_rows set is_locked = 1 where id=@.id
> >> > > > 3 return @.id
> >> > > > In different words : if in same time I call this procedure from 2
> >> > > > different connections, it will return 2 different record.
> >> > > > I also want that first call to procedure will not generate lock error
> >> > > > for second call in same time(just second call will wait for finish
> >> > > > first , it is ok.)
> >> > > > Thanks- Hide quoted text -
> >> > > - Show quoted text -
> >> > You wroute:
> >> > SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) WHERE is_locked
> >> > =0 )
> >> > -- Error checking
> >> > Could you explain what is reason for "Error checking " and what I can
> >> > do.
> >> > If some body in same time call same procedure , I can get error ?
> >> > Thanks- Hide quoted text -
> >> - Show quoted text -
> > Thanks,
> > it was very helpful. One last question:
> > if statetement "SELECT top 1 id FROM my_rows (UPDLOCK) ..." in second
> > call in store procedure can raise error because first call in first SP
> > still running (still locks the record) or it just will receive another
> > row?
> > Thanks- Hide quoted text -
> - Show quoted text -
In this case suggestion of John is not valid, because as I described
at start post,I need solution which will in every call to SP in same
time will return back a different row without any block.
Thanks|||On Wed, 26 Sep 2007 18:50:43 -0000, is_vlb50@.hotmail.com wrote:
>In this case suggestion of John is not valid, because as I described
>at start post,I need solution which will in every call to SP in same
>time will return back a different row without any block.
Hi is_vlb50,
As Tibor already mentioned, the READPAST hint can help you achieve what
you need. You'll find all the details in Books Online.
--
Hugo Kornelis, SQL Server MVP
My SQL Server blog: http://sqlblog.com/blogs/hugo_kornelis|||On Sep 27, 12:46 am, Hugo Kornelis
<h...@.perFact.REMOVETHIS.info.INVALID> wrote:
> On Wed, 26 Sep 2007 18:50:43 -0000, is_vl...@.hotmail.com wrote:
> >In this case suggestion of John is not valid, because as I described
> >at start post,I need solution which will in every call to SP in same
> >time will return back a different row without any block.
> Hi is_vlb50,
> As Tibor already mentioned, the READPAST hint can help you achieve what
> you need. You'll find all the details in Books Online.
> --
> Hugo Kornelis, SQL Server MVP
> My SQL Server blog:http://sqlblog.com/blogs/hugo_kornelis
the readpast applied only to update, delete, and writetext records.so
if i use
SELECT @.id = ( SELECT top 1 id FROM my_rows (readpast) WHERE
is_locked
=0 ) it will return same records to both simultanously call.
may be i can use it with UPDLOCK hint?
thanks|||Hi
You would be blocked whilst the second transaction finishes, hence the
possibility of a bottleneck. Why do you need to re-use the ids?
John
"is_vlb50@.hotmail.com" wrote:
> On Sep 26, 8:40 pm, "Tibor Karaszi"
> <tibor_please.no.email_kara...@.hotmail.nomail.com> wrote:
> > > if statetement "SELECT top 1 id FROM my_rows (UPDLOCK) ..." in second
> > > call in store procedure can raise error because first call in first SP
> > > still running (still locks the record) or it just will receive another
> > > row?
> >
> > TOP 1 can give you *any* row. If the row decided for is locked already by an incompatible lock, then
> > you will be blocked. If you want some other row, you might wan to check out the READPAST hint.
> >
> > --
> > Tibor Karaszi, SQL Server MVPhttp://www.karaszi.com/sqlserver/default.asphttp://sqlblog.com/blogs/tibor_karaszi
> >
> >
> >
> > <is_vl...@.hotmail.com> wrote in messagenews:1190831328.803233.199140@.19g2000hsx.googlegroups.com...
> > > On Sep 26, 6:02 pm, John Bell <jbellnewspo...@.hotmail.com> wrote:
> > >> Hi
> >
> > >> Readhttp://www.sommarskog.se/error-handling-I.htmlandhttp://www.sommarsko...
> > >> gives you a very good
> > >> idea of what to do. If you follow Erlands advice you would also have checked
> > >> for errors when you commit the transaction and have code to rollback.
> >
> > >> John
> >
> > >> "is_vl...@.hotmail.com" wrote:
> > >> > On Sep 26, 10:12 am, John Bell <jbellnewspo...@.hotmail.com> wrote:
> > >> > > Hi
> >
> > >> > > You select and update statement should be contained within one transaction
> > >> > > and you can use the UPDLOCK hint on the select statement to stop others
> > >> > > returning that value
> >
> > >> > > CREATE PROCEDURE GetLock ( @.id int OUTPUT ) AS
> > >> > > SET NOCOUNT ON
> > >> > > BEGIN TRANSACTION
> > >> > > SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) WHERE is_locked =0 )
> > >> > > -- Error checking
> > >> > > UPDATE my_rows SET is_locked = 1 WHERE id=@.id
> > >> > > -- Error checking
> > >> > > COMMIT TRANSACTION
> > >> > > RETURN
> >
> > >> > > Alternatively you could use
> >
> > >> > > UPDATE my_rows
> > >> > > SET is_locked = 1
> > >> > > WHERE id = (SELECT MAX(id) FROM my_rows WHERE is_locked = 0)
> >
> > >> > > but you would not know the ID that was updated.
> >
> > >> > > Using this type of locking can potentially cause a bottleneck and poor
> > >> > > performance.
> >
> > >> > > John
> >
> > >> > > "is_vl...@.hotmail.com" wrote:
> > >> > > > I need help with next problem:
> > >> > > > Exist table my_rows with next columns:
> > >> > > > id - int identity
> > >> > > > name - varchar
> > >> > > > is_locked - bit
> >
> > >> > > > I need create store procedure which will return every time diferent
> > >> > > > row for every request from different processes.
> >
> > >> > > > the pseudo code for the procedure :
> > >> > > > 1 select statement is :
> > >> > > > select @.id = selct top 1 id from my_rows where is_locked =0
> > >> > > > 2 update my_rows set is_locked = 1 where id=@.id
> > >> > > > 3 return @.id
> >
> > >> > > > In different words : if in same time I call this procedure from 2
> > >> > > > different connections, it will return 2 different record.
> > >> > > > I also want that first call to procedure will not generate lock error
> > >> > > > for second call in same time(just second call will wait for finish
> > >> > > > first , it is ok.)
> > >> > > > Thanks- Hide quoted text -
> >
> > >> > > - Show quoted text -
> >
> > >> > You wroute:
> > >> > SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) WHERE is_locked
> > >> > =0 )
> > >> > -- Error checking
> >
> > >> > Could you explain what is reason for "Error checking " and what I can
> > >> > do.
> > >> > If some body in same time call same procedure , I can get error ?
> > >> > Thanks- Hide quoted text -
> >
> > >> - Show quoted text -
> >
> > > Thanks,
> > > it was very helpful. One last question:
> > > if statetement "SELECT top 1 id FROM my_rows (UPDLOCK) ..." in second
> > > call in store procedure can raise error because first call in first SP
> > > still running (still locks the record) or it just will receive another
> > > row?
> > > Thanks- Hide quoted text -
> >
> > - Show quoted text -
> In this case suggestion of John is not valid, because as I described
> at start post,I need solution which will in every call to SP in same
> time will return back a different row without any block.
> Thanks
>|||On Wed, 26 Sep 2007 23:38:57 -0700, is_vlb50@.hotmail.com wrote:
>On Sep 27, 12:46 am, Hugo Kornelis
><h...@.perFact.REMOVETHIS.info.INVALID> wrote:
>> On Wed, 26 Sep 2007 18:50:43 -0000, is_vl...@.hotmail.com wrote:
>> >In this case suggestion of John is not valid, because as I described
>> >at start post,I need solution which will in every call to SP in same
>> >time will return back a different row without any block.
>> Hi is_vlb50,
>> As Tibor already mentioned, the READPAST hint can help you achieve what
>> you need. You'll find all the details in Books Online.
>> --
>> Hugo Kornelis, SQL Server MVP
>> My SQL Server blog:http://sqlblog.com/blogs/hugo_kornelis
>the readpast applied only to update, delete, and writetext records.so
>if i use
>SELECT @.id = ( SELECT top 1 id FROM my_rows (readpast) WHERE
>is_locked
>=0 ) it will return same records to both simultanously call.
>may be i can use it with UPDLOCK hint?
>thanks
>
Hi is_vlb50,
Yes, I thought John Bell already covered that.
First, you do a SELECT with UPDLOCK to make sure an exclusive lock is
acquired right away (to prevent two readers getting the same number),
*and* with READPAST to allow it to skip locked rows.
Then (in the same transaction), you do the update. Since you've already
got an exclusive lock, this will neven block or deadlock.
--
Hugo Kornelis, SQL Server MVP
My SQL Server blog: http://sqlblog.com/blogs/hugo_kornelis|||On Sep 28, 10:13 pm, Hugo Kornelis
<h...@.perFact.REMOVETHIS.info.INVALID> wrote:
> On Wed, 26 Sep 2007 23:38:57 -0700, is_vl...@.hotmail.com wrote:
> >On Sep 27, 12:46 am, Hugo Kornelis
> ><h...@.perFact.REMOVETHIS.info.INVALID> wrote:
> >> On Wed, 26 Sep 2007 18:50:43 -0000, is_vl...@.hotmail.com wrote:
> >> >In this case suggestion of John is not valid, because as I described
> >> >at start post,I need solution which will in every call to SP in same
> >> >time will return back a different row without any block.
> >> Hi is_vlb50,
> >> As Tibor already mentioned, the READPAST hint can help you achieve what
> >> you need. You'll find all the details in Books Online.
> >> --
> >> Hugo Kornelis, SQL Server MVP
> >> My SQL Server blog:http://sqlblog.com/blogs/hugo_kornelis
> >the readpast applied only to update, delete, and writetext records.so
> >if i use
> >SELECT @.id = ( SELECT top 1 id FROM my_rows (readpast) WHERE
> >is_locked
> >=0 ) it will return same records to both simultanously call.
> >may be i can use it with UPDLOCK hint?
> >thanks
> Hi is_vlb50,
> Yes, I thought John Bell already covered that.
> First, you do a SELECT with UPDLOCK to make sure an exclusive lock is
> acquired right away (to prevent two readers getting the same number),
> *and* with READPAST to allow it to skip locked rows.
> Then (in the same transaction), you do the update. Since you've already
> got an exclusive lock, this will neven block or deadlock.
> --
> Hugo Kornelis, SQL Server MVP
> My SQL Server blog:http://sqlblog.com/blogs/hugo_kornelis- Hide quoted text -
> - Show quoted text -
just for confirm final solution:
CREATE PROCEDURE GetLock ( @.id int OUTPUT ) AS
SET NOCOUNT ON
BEGIN TRANSACTION
SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) (READPAST) WHERE
is_locked =0 )
-- Error checking
UPDATE my_rows SET is_locked = 1 WHERE id=@.id
-- Error checking
COMMIT TRANSACTION
RETURN
Thanks|||Hi
From BOL:
WITH ( < table_hint > [ ,...n ] )
Specifies one or more table hints. For more information about table hints,
see FROM.
John
"is_vlb50@.hotmail.com" wrote:
> On Sep 28, 10:13 pm, Hugo Kornelis
> <h...@.perFact.REMOVETHIS.info.INVALID> wrote:
> > On Wed, 26 Sep 2007 23:38:57 -0700, is_vl...@.hotmail.com wrote:
> > >On Sep 27, 12:46 am, Hugo Kornelis
> > ><h...@.perFact.REMOVETHIS.info.INVALID> wrote:
> > >> On Wed, 26 Sep 2007 18:50:43 -0000, is_vl...@.hotmail.com wrote:
> > >> >In this case suggestion of John is not valid, because as I described
> > >> >at start post,I need solution which will in every call to SP in same
> > >> >time will return back a different row without any block.
> >
> > >> Hi is_vlb50,
> >
> > >> As Tibor already mentioned, the READPAST hint can help you achieve what
> > >> you need. You'll find all the details in Books Online.
> >
> > >> --
> > >> Hugo Kornelis, SQL Server MVP
> > >> My SQL Server blog:http://sqlblog.com/blogs/hugo_kornelis
> > >the readpast applied only to update, delete, and writetext records.so
> > >if i use
> > >SELECT @.id = ( SELECT top 1 id FROM my_rows (readpast) WHERE
> > >is_locked
> > >=0 ) it will return same records to both simultanously call.
> > >may be i can use it with UPDLOCK hint?
> > >thanks
> >
> > Hi is_vlb50,
> >
> > Yes, I thought John Bell already covered that.
> >
> > First, you do a SELECT with UPDLOCK to make sure an exclusive lock is
> > acquired right away (to prevent two readers getting the same number),
> > *and* with READPAST to allow it to skip locked rows.
> >
> > Then (in the same transaction), you do the update. Since you've already
> > got an exclusive lock, this will neven block or deadlock.
> >
> > --
> > Hugo Kornelis, SQL Server MVP
> > My SQL Server blog:http://sqlblog.com/blogs/hugo_kornelis- Hide quoted text -
> >
> > - Show quoted text -
> just for confirm final solution:
> CREATE PROCEDURE GetLock ( @.id int OUTPUT ) AS
> SET NOCOUNT ON
> BEGIN TRANSACTION
> SELECT @.id = ( SELECT top 1 id FROM my_rows (UPDLOCK) (READPAST) WHERE
> is_locked =0 )
> -- Error checking
> UPDATE my_rows SET is_locked = 1 WHERE id=@.id
> -- Error checking
> COMMIT TRANSACTION
> RETURN
> Thanks
>
No comments:
Post a Comment