Top 5 features you would like to see in the next version of SQL Server

  • Article

Here is an opportunity to discuss about the top 5 features that you would like to see in the next version of SQL Server. It would be nice if you can include a sentence for each feature explaining why it is required. Please use features of SQL Server 2005 as starting point meaning do not ask for those which are already in SQL Server 2005. And the features should be restricted to the query language (SQL) and the programming language (TSQL) only. Examples include enhancements to DDLs, new relational features, new programming capabilities or data types.

Lastly, note that this is just an informal request from me. This doesn't mean that these features will get into the next version of SQL Server. But I will take these requests and file them along with existing suggestions/sqlwish emails that many in our team maintain. And it can help prioritize the list for the product eventually.

Thanks
Umachandar

--

Summary of the responses received so far can be found at https://umachandar.members.winisp.net/files/top_5_features_summary.htm

Comments

  • Anonymous
    October 07, 2005
    I'm afraid I don't know if this is already in sql 2005, but the very top item on my list is to get rid of the stupid "This might cause cycles or multiple cascade paths" error message when trying to add a foreign key constraint.

    Not sure if that counts as part of SQL / TSQL or not (it's an error message that shows up when you do ALTER TABLE but the language can't do much if the engine doesn't support it I guess).

    Not supporting cascade cycles I can handle, but multiple acyclic cascade paths is a very common requirement in what I do and I'm always stuck implementing poor, inferior, hacky workarounds, or just punting on the issue entirely and leaving possible scenarios entirely unhandled.
  • Anonymous
    October 07, 2005
    Yes, ALTER TABLE definitely counts or in this case creation of any FK constraint that can introduce cycles. Unfortunately this is not fixed in SQL Server 2005 also. We only added the SET NULL and SET DEFAULT options for the cascading constraints.
  • Anonymous
    October 07, 2005
    I too am sorry that I am not as familiar with SQL 2005, but here goes...

    * I would like to segment db objects by functional group, much the same way you can do owner.name (dbo.Authors), I would like code_segment.name.

    * Have FTS more integrated into the table and column structure so that the queries will run against a regular index or FTS index to produce the quickest result via the optimizer all transparent to the user.
  • Anonymous
    October 07, 2005
    I know it's simple, but it's the one thing I really miss from other DBs that makes paging through data so much easier: the LIMIT clause.
  • Anonymous
    October 07, 2005
    Re: "would like to segment db"

    Please look at the new CREATE SCHEMA feature. You can create schemas that correspond to your functional groups. See also: the data model in the AdventureWorks database.

    Let us know what you think and how that works for you.

    Clifford Dibble,
    SQL Server Engine, Program Manager
  • Anonymous
    October 07, 2005
    The comment has been removed
  • Anonymous
    October 07, 2005
    I hate this error message:

    Conversion failed when converting a value of type varchar to type int. Ensure that all values of the expression being converted can be converted to the target type, or modify query to avoid this type conversion.

    You get this when trying to alter a table converting the data or when you are inserting from one table to another. The only clue that I have as to which column is causing the problem is that I know it is a varchar column going into an int.

    I see this all the time when loading and converting questionable data. It would help greatly to know the name of the column. It would be nice if it could give the row or PK value for the row causing the problem as well.
  • Anonymous
    October 08, 2005
    The comment has been removed
  • Anonymous
    October 08, 2005
    Got it. This is just one of the several restrictions we have on cascading relationships.
    Regarding your comment about unique index allowing multiple NULLs, it is a valid request. ANSI SQL standard which doesn't cover indexes but talks about UNIQUE constraints that can allow only one NULL or multiple NULLs. They leave it upto the implementation and in fact the one NULL implementation is considered entry level compliance for this feature.
    You can however use an indexed view in SQL Server 2000/2005 to workaround this limitation. This provides similar behavior in that the data in the table will follow the rules of the constraint. Here is a simple example:

    set nocount on;
    use tempdb;
    go
    create table T (
    i int not null primary key
    , j int null /* need to make this unique only for non-null values. */
    );
    go
    create view T_j
    with schemabinding
    as
    select j
    from dbo.T
    where j is not null;
    go
    create unique clustered index UQ_T_j on T_j(j);
    go
    insert into T values(1,1);
    insert into T values(2,NULL);
    insert into T values(3,NULL);
    insert into T values(4,1); -- this will fail
    go
    select * from T;
    go
    drop view T_j;
    go
    drop table T;
    go

    --
    Umachandar
  • Anonymous
    October 08, 2005
    The comment has been removed
  • Anonymous
    October 08, 2005
    DOMAIN support.

    That's all :)

    Thank you.
  • Anonymous
    October 09, 2005
    Real date and time types, for natural separation of the two.
  • Anonymous
    October 10, 2005
    SAVING QUERY RESULTS:
    I would like to see more choices for saving results in Query Analyzer. For instance it would be nice to save the grid results directly to an .xls, .html file etc.

    Currently you have to save the results as a text file via menu Results in Text, or File which is limited import the results to Excel.

    So how about saving the results as insert statements also?

    DDL GENERATION:
    Is there a way to reverse engineer a dbms schema in Enterprise Manager, or Query Analyzer? It would be nice to generate the ddl of a selected dbms.


    NAMING STANDARDS:
    I have noticed that specific terms change in the dbms and in the books online (updated).

    In Enterprise Manager you see the term Role, but if you run:

    EXEC sp_helpuser you see group name listed. In the books online (updated) it lists:

    GroupName sysname Roles to which UserName belongs.
    Group_name sysname Name of the role in the current database.

    So why are the terms role and group used interchangeably? This is confusing.

    Thanks,

    Mark P
  • Anonymous
    October 10, 2005
    Row value constructors, which would in turn let me do: WHERE (a, b) IN (SELECT c, d FROM e)
  • Anonymous
    October 12, 2005
    Row-value constructors

    ANSI-compliant UNIQUE indexes/constraints (allows multiple NULLs)

    Native DATE and TIME types

    Autonomous transactions - the ability to commit data in a nested transaction outside the scope of an outer transaction

    Deferrable constraints

  • Anonymous
    October 12, 2005
    The comment has been removed
  • Anonymous
    October 12, 2005
  1. DATE and TIME, compatible with existing DATETIME, not the mess that was in Beta 1.

    2) ALTER TABLE to insert a colunm at any place, or move a column.

    3) A function that returns the full call-stack, @@procid is not enough.

    4) SET STRICT ON, under which it's an error to refer a table that does not exist, under which many implicit conversions are disallowed, calls to stored procedures are checked for matching parameters stc.

    5) Deferred constraints and triggers.

    6) UPSERT/MERGE command.
  1. Separation of DATE and TIME types and support for temporal features

    2) Regular Expressions

    3) Row-value constructors

    4) Full support of OVER clause for both ranking functions and aggregates:
    PARTITION BY
    ORDER BY
    ROWS

    5) MERGE / UPSERT
  • Anonymous
    October 13, 2005
    I have not come across any database having this feature . This may be difficult to implement if not impossible

    Need to run through all SPs and provide table names and columns used in where, order by , group by for each SP. This is needed for impact analysis and performance tuning .
  • Anonymous
    October 13, 2005
    I'll have to add a +1 on the deferred constraints.
  • Anonymous
    October 13, 2005
    The comment has been removed
  • Anonymous
    October 13, 2005
    Another one;)

    Expose Extended Properties through updateble system views instead of system functions.

    and another (which should be done anyway):

    The INFORMATION_SCHEMA views should be updated. SQL92 doesn't exists anymore...its been replaced by SQL99. Those views should at least be SQL99 compliant. SQL2003 would be even better.
  • Anonymous
    October 13, 2005
    My Fridays' desired features are: 1) T-SQL arrays (I don´t want to implement them using SQLCLR code :-)) 2) Restore of tables by their names, sort of RESTORE TABLE FROM (I don't want to organize my table on filegroups to implement it). 3) TempDb not a global resource, TempDb for each database if I need a db with intensive tempdb usage.

    Best regards!
    ~gux

  • Anonymous
    October 13, 2005
  1. I'd like to add a vote for Itzik's native support for regular expressions in T-SQL, if that's what he meant.
    2. RESTORE DATABASE ... WITH SHRINK to <size>
    3. Ability to freeze/bind plans
    4. Recycle bin by default, similar to Oracle's Flashback technology
    5. Logical DB mirroring (right now DB mirroring is only physical, i.e. you can't generate SQL statements from the log records and apply the SQL statements instead)

    Linchi
  • Anonymous
    October 13, 2005
    The top 5 features:

    (A) Support intellisense in SSMS. Please!
    (B) A nice GUI for Service Broker Management from SSMS.
    (C) Separate DATE and TIME data types.
    (D) HTTP support in Service Broker messaging so that we do not have to play around with TCP.
    (E) Ability to read mails using the new DBMail feature.
  • Anonymous
    October 14, 2005
  1. More analitical functions (like Oracle)
    2) Use of external certificates or smartcards with data encryprion
    3) Arrays
    4) Intellisense
    5) Ability that user can build his/her navigation tree with desired DB objects. I have problem with browsing stored procedures because I have more than 3K SPs in DB, and I want to work with only few SPs and tables.
  • Anonymous
    October 14, 2005
  1. Date and Time separate data types

    2) Duration data type for date / time arithmetic, e.g. "two days" not "September 18

    3) REGEX capability within T-SQL

    4) ANSI std row constructor

    5) For me the most exciting, but perhaps pie-in-the-sky, request: even more fully object/relational integration. CLR inside SQL Server is a wonderful start; it'd be really incredible to get to a full object-relational model in the future.
  • Anonymous
    October 14, 2005
    The comment has been removed
  • Anonymous
    October 14, 2005
    I'll echo some other's requests:

    1. CREATE DOMAIN
    2. NATURAL JOIN
    3. Queries in CHECK constraints
    4. Dynamic PIVOT
    5. Database-specific error messages

    And for a real pie-in-the-sky request, I'd love to be able to define custom errors for constraint violations:

    ALTER TABLE myTable ADD CONSTRAINT CHK_Amount CHECK(Amount BETWEEN 1 AND 10) RAISERROR('Amount must be between 1 and 10.', 16, 1)

    Of course I could also ask for TRY...CATCH support on constraint violations, but that's pushing it. :)

    Thanks!
  • Anonymous
    October 14, 2005
    I already one added one too many, but I forgot a very important
    thing, which makes about #2 on my wishlist after date/time:
    Being able to pass table variables as parameters. Suggestively you can define a table type, and then you can declare tables and parameters of that type. This is an essential feature to make i tpossible to pass tabular data between stored procedure between stored procedures.

  • Anonymous
    October 14, 2005
    Automatic promotion of a secondary standby database to the primary database, and vice versa.
  • Anonymous
    October 14, 2005
    Regarding Gustavo Larriera's comment:

    You can freeze/bind plans for specific queries in SQL Server 2005 using plan guides (sp_create_plan_guide) and plan forcing (the USE PLAN <xml_plan> query hint).
  • Anonymous
    October 16, 2005
    Thanks for asking UC :-)

    Separate date and time datatypes. Same basic semantics and rules as current ones.

    Row value constructors

    Some stricter level for implicit datatype conversion. Say you define a parm as nvarchar and use parm in WHERE and col is varchar. No SARG. And this doesn't only have perf implications. Often it is a mistake to use different datatypes, and I want to get help finding those mistakes.

    A function to return datatype from an expression. IsWhatDatatype(expression) can return 'int', 'datetime', etc. Expression can be constant, column name, variable, literal etc. Engine has rules for this, why not expose these rules? Do all reader here know what datatypes below are of?
    1
    3422434
    1.23
    'Hello'

    DOMAIN as per ANSI SQL. If not doable, I suggest SIMILAR predicate (reg exp).
  • Anonymous
    October 17, 2005
    The way the table datatypes are implemented needs to be expanded so that it is possible to define the table datatype as a replica of an existing table. The sps become really clumsy with all the definitions.

    The second thing is respect to differed resolution. The thing differed resolution is not necessarily a good thing. Atleast there should be a way to enforce differed resolution by means of a switch. I know we do have sp_dbcompatibilitylevel. I am not sure if that is the right way to use. We have faced numerous problems in Production because of this.
  • Anonymous
    October 21, 2005
    I think allowing to insert an identity column on the fly should be a good enhancement.Right now as an alter statement we cannot add identity , although it allows through Enterprise.
  • Anonymous
    October 24, 2005
    Hi Umachandar

    1. CREATE DOMAIN and ALTER DOMAIN (without losing existing data)
    2. delay enforcing of foreign key constraints until COMMIT (deffered constraints)
    3. Reporting Services: KeepTogether flag on subreports AND textboxes (so you can turn it off)
    4. Report Builder: allow other data source than SQL Server/Analysis Services (esp. own providers)
    5. Profiler: allow sorting by clicking on columns (like in VS and Outlook)
  • Anonymous
    October 24, 2005
    The comment has been removed
  • Anonymous
    October 24, 2005
  1. Native date and time, to support milliseconds with 1ms as accuracy.
    2. Regular expressions.
    3. Intellisense.
    4. Array datatype.
    5. Easier ways to locate table scans.
  • Anonymous
    October 25, 2005
    It would be interesting to see that results summary ordered by total score (either just ordered by count or ordered by sum(6 minus preference)).

    I know that this poll doesn't actually mean anything as far as what will be implemented but it would be nice to be able get an idea of which features "won" ;)
  • Anonymous
    October 27, 2005
    I'll add another vote to deferrable constraints.
  • Anonymous
    October 27, 2005
    Stuart,
    Good idea! I sorted the features by count instead of category. I also moved the tools and other requests to a different table since that is not of interest to me (I have however forwarded some of the requests to people in the other teams).
    And don't underestimate the importance of getting these results. Lot of people who are involved in the planning of features for next release of SQL Server are aware of the list. It will help in strengthening some of the existing requests and provide additional validation. So don't be surprised when you see some items from here make it to a future version of SQL Server.

    --
    Umachandar
  • Anonymous
    October 27, 2005
    Well, I am so busy trying to understand all the new features that are there that I'm not ready to say what I want next (except I want the migration tools to deal with logins from prior versions). But I'll repeat what I said on the beta news group: what I want more than anything is to not have the next release be 5 years away. This release is huge -- and I have been working with it since beta 1. I think in some ways it is worse than the 6.x - 7 changes in terms of the learning curve. Many of the changes are far more subtle, which makes them a) harder to discover and b) harder to understand

    Sharon
  • Anonymous
    October 27, 2005
    I would like folders for DTS packages, i have too many of them and very hard to keep track of.
  • Anonymous
    October 31, 2005
    The comment has been removed
  • Anonymous
    November 07, 2005
  1. ability to chose temporal/lossless storage. (Create database mydb with storage=lossless)
    2. Unique Transaction stamps akin to timestamps but identical per transaction.
    3. query enhancements for lossless schemas whereby one might ask:
    select ... from ... join ... join ... where ... and DataBasePointInTime = dateadd(Y,-1,getutcdate())
  • Anonymous
    November 09, 2005
  1. Variable table and field names in select statements, etc.

    e.g.:
    declare @theTable varchar(50) -- (or some other data type like TableName)
    declare @aField varchar(50) -- (or datatype FieldName)
    set @theTable = 'Customers'
    set @aField = 'LastName'
    select @aField from @theTable

    2. A way to return a range of rows in the middle of a query (like TOP 10, but the 6th group of 10)

    e.g. SELECT ROWS 51 TO 60 FROM Customers ORDER BY LastName.
  • Anonymous
    November 10, 2005
    Ability to use IN with a variable.

    e.g.

    SET @Pets = '(''cat'', ''dog'', ''fish'')'
    SELECT * FROM animals WHERE species IN @Pets

    To give the ability to pass a list in to a stored procedure in a parameter.
  • Anonymous
    November 10, 2005
  1. The Ability to work with arrays in Stored Procs in T-SQL including being able to pass them in as params via ADO
    2) CREATE ASSERTION
    3) Real Temporal support
    4) Intellisense in tools
    5) A Whitehorse based IDE for DB modeling instead of Visio or any other stop gap diagraming solution

    ps great job on 2005!
  • Anonymous
    November 15, 2005
  1. Intellisense
    2) The ability to track what procedures are run - How often / Last run / etc - Pain in the nexk to find obsolete SPs
    3) UPSERT
    4) Pivot table with non specified values - ie from a select statement and not supplied using in and hard coding the values
    5) Natural Join
  • Anonymous
    December 21, 2005
    The comment has been removed
  • Anonymous
    December 21, 2005
    Oh, and I would also like to vote on Intellisense, also inside the Management Studio. You can put that in position 4.

    So:
    1) Table names as variables in T-SQL queries
    2) Member functions
    3) Peer connections
    4) Intellisense
    5) Table Folders

    Non T-SQL:
    6) LINQ in C# stored procedures
    7) More functionality added to SHFT-CTRL-M
  • Anonymous
    January 04, 2006
    Using pooled Windows Authentication. Oracle does this with their Proxy Authentication feature.
  • Anonymous
    January 07, 2006
  1. How about before and after triggers (both row and statement if possible), and allowing to change the inserted table in before triggers
    2) some equivalent of Oracle's autonomous transactions
    3) MERGE statement
    4) automatic audit columns
    5) allowing query in check constraints
  • Anonymous
    January 09, 2006
    I would like see "Intellisense" in my SQL Server Managment Studio, the same way i have in Visual Studio.
  • Anonymous
    January 14, 2006
    I agree with the following:

    1) Table names as variables in T-SQL queries
    2) Member functions
    3) Peer connections
    4) Intellisense
    5) Table Folders

    With regards
    Borum.NET
  • Anonymous
    January 23, 2006
    I would love to see cross database foreign keys. As larger clients start to use sql server it becomes more important to break the data down into manageable db sizes. I want to have an orders db and a customer db and not have to write triggers to make sure I can't delete a customer that has orders in the orderdb.
  • Anonymous
    February 23, 2006
    The comment has been removed
  • Anonymous
    February 23, 2006
    I have a top two :-) (although not t-sql related)
    * help 'in' the management studio
    * IntelliSense when writing queries
    Kind regards.
  • Anonymous
    March 05, 2006
    From a data warehousing perspective, the following additionals in the relational engine would provide great benefits.
    1) Bitmap indexes
    2) UPSERT / MERGE statement
  • Anonymous
    March 09, 2006
    Here is my request about a future release of SQL Server :

    1) date / time types
    DATE type
    TIME type
    INTERVAL type
    DATETIME that accuracy is 0.001 second and include TIMEZONE
    and all operators to do with (like '2005-01-01' + 6 MONTH)

    2) UNIQUE ISO SQL constraint (accept multiple NULLs)

    3) ROW VALUE CONSTRUCTOR

    4) Deffered constraints

    5) ARRAY datatype with multisets operators (MULTISET, COLLECT, UNNEST [WITH ORDINALITY], CARDINALITY, ELEMENT, IS A [NOT] SET, UNION, INTERSECT, EXCEPT, FUSION...)

    6) BEFORE Triggers

    7) SIMILAR predicate (SQL:1999 regular expression)

    8) SQL ASSERTION

    9) MERGE statement (SQL:1999 UPDATE/INSERT)

    10) NATURAL JOIN

    11) RETURN NULL ON NULL INPUT statement in stored procedure and UDF

    I read some demands wich I desagree completely :
    Dynamic Pivot :
    I am completly afraid to see such a non relational operator in a great RDBMS. Microsoft would be condamned for that... So Dynamic Pivot is a pure nightmare !
    Use DMX queries !

    Variable table and field names in select statements :
    This is also a pure nightmare. Object's name cannot be a variable...

    ALTER TABLE to insert a colunm at any place, or move a column :
    This is an anti relationnal feature. There is absolutly no order at any place in a RDBMS. Only the order you set it. Remember that a table is a bag of marble. ISO SQL make ordinal position of column only to simplify INSERT statement and this is a mess ! Also a mess the ability to specify the ordinal position of the expression in the clause SELECT of the SELECT statement for the ORDER BY clause. Thoses two SQL ISO features are stupids.

    And some other that is not necessary :

    paging through data : easy in 2005 by the use of ROW_NUMBER()

    DOMAIN : can be done by sp_addtype, CREATE RULE, CREATE DEFAULT, sp_bindrule, sp_bindefault...
    wich I think is much more interresting in a matter of chaching rules

    A +
  • Anonymous
    March 16, 2006
    stored procs and functions should accept table variable parameters... SQL needs this for Reporting Services reports to be able to use multivalued parameters with stored procs with out kludgy string split functions in the stored procs
  • Anonymous
    March 23, 2006
    The comment has been removed
  • Anonymous
    March 23, 2006
    There is only one I want to add that hasnt already been mentioned.

    The ability to get a native sql statement for a smdl query.

    The report builder available for 2005 is incredibly useful, but could be tremdously valuable for people wanting to integrate the resulting datasets with other applications (MS, .Net, ODBC, or otherwise).

    See this google posting...
    http://groups.google.com/group/ReportingServices/browse_thread/thread/533350f97660cc20/1842f4d794c839b6%231842f4d794c839b6
  • Anonymous
    March 30, 2006
    Better statistics, ability to create larger histograms to get a more accurate representation of the data in a table, applies to very large tables.
  • Anonymous
    April 28, 2006
    LIMIT LIMIT LIMIT

    ROW_NUMBER is a hassle to code in web applications.  It's that simple.
  • Anonymous
    May 17, 2006
    Ability to easily measure Server usage @ DB Level from a performance, memory, utilisation point of view without having to use Profiler and Analyse. This is important in consolidated environments
  • Anonymous
    May 26, 2006
  1.  A way to find the list of all table/procedures names that a stored procedure is planning to use "deferred name resolution" on.  That way we can at lease process the list to see if all of them are ok, or there are ones we should look at.

    2.  Indexes on functions of columns.  That way, when you have a several million row table with an index varchar column and you are querring the table with a parameterized querry that is using a Nvarchar parameter, it will not do an index scan and conversion on all the rows for each querry.  (Third Party Apps.)

    Ryan M. Hager

    (r) hager (A)(T) hager1 (D) (O) (T) Com with no "(",")", or spaces and AT=@
  • Anonymous
    June 02, 2006
    FK across DB's
  • Anonymous
    June 04, 2006
    The comment has been removed
  • Anonymous
    June 10, 2006
    I would suggest your team should look into discussion-goups for the most repeated questions (pattern study) and fix those ASAP.

    Examples in BOL are pathetic, those needs a real improvement.  If you guys DO NOT want to write any examples then you should tei-up with some partner who have good examples on their site :)
  • Anonymous
    June 26, 2006
    Support cascading cyclic and multipath FK-references...
    please... :P
  • Anonymous
    June 28, 2006
    I vote for the LIMIT keyword, it would make things much easier...
  • Anonymous
    July 12, 2006
    The comment has been removed
  • Anonymous
    July 12, 2006
    Allow expression list as a valid expression.  This allows for non-scalar subquery comparisons

    ... WHERE (a,b) IN (SELECT x,y)

    I gather this is in the official SQL standard (ISO/IEC 9075-2:2003 section 8.4)

    There really seems to be no workaround to not having this when as bind variables tuples are provided:
     
       WHERE (a,b) IN ((1,2), (1,3))

    Oracle has supported this atleast since version 7, so I expect most DBs already have this functionality.
  • Anonymous
    September 07, 2006
    Better date time formatting options. Rather than the numbers in the convert statement allow passing a mask to format date times. eg select convert(varchar,getdate(),"dd/m/yy hh:mm")
  • Anonymous
    September 09, 2006
    why this stored procedure is not working plz tell me correct code---------------------------
    -----------------------------------------------
    -------------run this code --------------------

    use northwind
    go
    -----------table created----------

    create table companey

    (
       compname varchar(20),
       price  integer
    );
    go


    -----------value inserted---------------

    insert into companey values('sskEnterprises', 200000);
    insert into companey values('sskGoods', 100000);
    insert into companey values('sskBaveries', 14700000);
    insert into companey values('sskFoods', 4100000);
    insert into companey values('mhkEnterprises', 74500000);
    insert into companey values('mhkfoods', 145000);
    insert into companey values('mhkgoods', 16000);
    insert into companey values('sskshipping', 145800);
    go


    -----------------------check the value----

    select * from companey where compname like 'mhk%';



    go

    --------------------procedure created-----------
    create procedure sp_like

    @cn varchar(20)
    as

    select * from companey where companey.compname like @cn

    go


    -------------- procedure droped---------------

    drop procedure sp_like

    go

    -----------------procedure executed-----------
    exec sp_like mhk;



  • Anonymous
    October 03, 2006
    A more intuituve way than CTE's for hierarchial queries, something like Oracle's "CONNECT BY".
  • Anonymous
    April 26, 2007
  1. MPP architecture
  2. Populate tables with test data (words or numbers which mean nothing)
  3. Upsert command
  4. Enable SCD type 3 on SSIS' SCD Wizard.
  5. Bitmap index
  6. Perl-like regex string manipulation
  7. Oracle-like to_date function
  8. Like bulk insert but to export data out (not OS command like bcp -out but a standard T-SQL command)

  • Anonymous
    May 31, 2008
    Here is an opportunity to discuss about the top 5 features that you would like to see in the next version of SQL Server. It would be nice if you can include a sentence for each feature explaining why it is required. Please use features of SQL Server 200

  • Anonymous
    June 05, 2008
    Here is an opportunity to discuss about the top 5 features that you would like to see in the next version of SQL Server. It would be nice if you can include a sentence for each feature explaining why it is required. Please use features of SQL Server 200