| Author |
Topic |
|
spirit1
Cybernetic Yak Master
11752 Posts |
 Posted - 2006-02-24 : 08:22:52
|
we're having a Slovenian dev user group meeting on March 6th and we're having these 2 guest speakers:quote:
Don Vilen is on the SQL Server team in Redmond, with eleven years of SQL Server experience at Microsoft. For the past five years Don has been a Program Manager on the Storage Engine component of SQL Server, focusing on high availability and the storage of large objects. For the preceding four years Don created and delivered training for Microsoft's SQL Server support engineers in EMEA and around the world. Prior to joining Microsoft Don was on the computer science faculty at a California university and was also the Vice President of Software Development for a software company. Donald Farmer is Group Program Manager for the SQL Server Integration Services team. Donald joined Microsoft in July 2001. Moving over from Scotland, he is one of the few incomers to find Seattle warmer and drier than he is used to! Donald started at Microsoft in the Analysis Services team, but quickly moved to DTS, as it was then called. Before Microsoft he worked on software projects as diverse as desktop publishing, hydrographic modeling, document analysis for medieval history, management systems for fish farms and a rapid development and deployment tool for data warehouses.
topic is SQL server 2005any questions anyone would like me to ask and report back?Go with the flow & have fun! Else fight the flow  |
|
|
X002548
Not Just a Number
15586 Posts |
|
|
spirit1
Cybernetic Yak Master
11752 Posts |
Posted - 2006-02-24 : 10:09:57
|
lolok  Go with the flow & have fun! Else fight the flow |
 |
|
|
Kristen
Test
22859 Posts |
Posted - 2006-02-24 : 12:57:10
|
| Not enough DATEDIFF time to fix it? Kristen |
|
|
|
jhermiz
3564 Posts |
Posted - 2006-02-24 : 15:50:47
|
I had an MS article a few weeks back that stated they did not implement the date vs datetime functionality because doing so affected various other capabilities in sql server 2005. I'm trying to find out where I found that article since its been some time, but basically they said doing so would delay the product even more and add on to a lot more complexity to changing other pieces of code.So I guess that mean's well be seeing "How do I get rid of the time portion from my date" question for another 10 more years :).
Keeping the web experience alive -- [url]http://www.web-impulse.com[/url]RS Blog -- [url]http://weblogs.sqlteam.com/jhermiz[/url] |
|
|
|
Kristen
Test
22859 Posts |
Posted - 2006-02-25 : 01:37:16
|
| Even faking it with a pseudo DATE time that did the behind-the-scene strip of TIME would have done for me!Or DATE(MyDateTime) and TIME(MyDateTime) fucntions ...... perhaps that's the solution I should offer, instead of trying to explain why DATEADD(Day, DATEDIFF(Day ...)) is better than CONVERT(datetime-to-varchar-and-back-again)!Kristen |
|
|
|
Kristen
Test
22859 Posts |
|
|
jen
Master Smack Fu Yak Hacker
4110 Posts |
Posted - 2006-03-01 : 00:36:21
|
| request for a summary of the changes they made from sql2000 to sql2005it's helpful if we're already aware of these things rather than "discover" them on our own and encounter them when we roll out or try the product eh? TIA--------------------keeping it simple... |
|
|
|
mmarovic
Aged Yak Warrior
518 Posts |
Posted - 2006-03-02 : 12:13:22
|
| Do they plan to implement row versioning any time soon (if at all)? |
|
|
|
Michael Valentine Jones
Yak DBA Kernel (pronounced Colonel)
7020 Posts |
Posted - 2006-03-02 : 14:26:33
|
quote:
Originally posted by mmarovic
Do they plan to implement row versioning any time soon (if at all)?
What is row versioning?I assume you mean something different than the ROWVERSION (timestamp) datatype, which is available in SQL 2000.CODO ERGO SUM |
|
|
|
mmarovic
Aged Yak Warrior
518 Posts |
Posted - 2006-03-02 : 15:17:58
|
| This is the postgreSQL feature (not sure about Ingres). Internally, each change in db is written as a new row. So, when you need to have consistent view of data at the time read started you can do it without locking. Oracle has the same ability but it uses transaction log to achieve that goal. |
|
|
|
spirit1
Cybernetic Yak Master
11752 Posts |
Posted - 2006-03-06 : 16:12:31
|
ok got back: so while it's fresh in my head:1. mmarovic: forgot to ask about rowversioning.. sorry...  2. brett: becaues you can do stuff like this with CLR so there is no need to implement time and date as real sql datatypes. they figured it would be better to make other stuff better.3. jen: there's no brief summary in one place . it would be huge. you can go to MS sql server site and get a list of new features.they told us that importing from a flat file into the DB is now faster using integration services than using BCP with EDIT: got it: it's FastParse (http://msdn2.microsoft.com/en-us/library/ms139833.aspx). the improvement is 7-20%. that's because there's now a library that performs superfast standard datatypes conversion from char in SSIS.- if you'r db has n filegroups you can still access the db if the primary filegroup isn't corrupt. of course you can't access the data in corrupt filegroups.- there's index disabling. you can create an index say for monthly report. run the report. disable the index if you don't need it for other stuff, when the index gets enabled it's rebuilt.- there's no tool for watching TLOGS. that's reserved to 3rd party becaues MS sells the format of the TLOG (that's how i understood it). You can save TLOG to txt file and try parsing it from there. but since replication works from reading the TLOG i guess one could hack something up...- vertical partitioning is deprecated you can say which data will be in what file group. it's supposed to be the preffered way of storing large amounts of data in one table... there was no example but that's how i understood it.- there's now statement recompile instead of batch recompile.so when running a sproc we get to statment that needs to be recompiled only that statement is recompiled and [EDIT]NOT[/EDIT] all the statements from that point on. EDIT: that detail is an important part i missed - full text now builds 50x faser. queries with FT are executed 10-100 times faster.those are some of the points i remember.also got a VS2005 and SS2005 Standard editions. Go with the flow & have fun! Else fight the flow |
|
|
|
jen
Master Smack Fu Yak Hacker
4110 Posts |
Posted - 2006-03-06 : 20:01:20
|
wow great Mladen,now you can Test those stuff they mentioned  are there really that much differences? Well, that means I need to be not-lazy for a couple of months to properly evaluate the productthanks!--------------------keeping it simple... |
|
|
|
Kristen
Test
22859 Posts |
Posted - 2006-03-07 : 00:48:48
|
| "2. brett: becaues you can do stuff like this with CLR so there is no need to implement time and date as real sql datatypes. they figured it would be better to make other stuff better."How crap is that ...What I read was they did it in CLR, instead of core, and it had some problems so they chopped it. And now they want US to do it in CLR eh? Kristen |
|
|
|
Kristen
Test
22859 Posts |
Posted - 2006-03-07 : 04:49:42
|
| Spirit1: Did you delete a reply? "Active Topics" is saying you replied at 04:40:31, but I'm only seeing my reply at 00:48:48 :-( |
|
|
|
spirit1
Cybernetic Yak Master
11752 Posts |
Posted - 2006-03-07 : 04:58:42
|
| no i edited my last reply prior to this one.i forgot one "NOT" that changes the meaning of the statement Go with the flow & have fun! Else fight the flow |
|
|
|
spirit1
Cybernetic Yak Master
11752 Posts |
Posted - 2006-03-07 : 05:01:13
|
| oh yeah... i also asked for a feature i want a culture inspecific char for a decimal point.Go with the flow & have fun! Else fight the flow |
|
|
|
spirit1
Cybernetic Yak Master
11752 Posts |
Posted - 2006-03-07 : 06:27:11
|
| one more thing i remembered:there is no upsert statment planned.Go with the flow & have fun! Else fight the flow |
|
|
|
Frank Kalis
Constraint Violating Yak Guru
413 Posts |
Posted - 2006-03-08 : 02:59:17
|
quote:
Originally posted by mmarovic
This is the postgreSQL feature (not sure about Ingres). Internally, each change in db is written as a new row. So, when you need to have consistent view of data at the time read started you can do it without locking. Oracle has the same ability but it uses transaction log to achieve that goal.
Mirko, do you mean Oracle's FLASHBACK DATABASE feature?If so, I might be wrong, but I haven't heard anything about it yet.--Frank KalisMicrosoft SQL Server MVPhttp://www.insidesql.deHeute schon gebloggt? http://www.insidesql.de/blogs |
|
|
|
mmarovic
Aged Yak Warrior
518 Posts |
Posted - 2006-03-08 : 03:21:12
|
| No, it is an old Oracle functionality. Oracle uses rollback segements and combine it with what was read from data pages to offer consistent view of data at the time transaction started (if I remember well, I havent't worked on Oracle either for some time). |
|
|
|
|
questions for SQL server 2005 dev team members