« first day (2898 days earlier)      last day (1955 days later) » 

6:35 AM
@McNets Yes. What I thought too.
Morning
 
7:11 AM
Please consider merging your accounts (1 & 2), so that you can edit your own questions directly and consolidate your reputation. Have a look at I accidentally created two accounts; how do I merge them? (Meta | Help Center) for more information. — hot2use 46 secs ago
 
@hot2use one of them it's unregistered, I don't know it them can be merged.
Morning
 
7:53 AM
Morning
 
Morning
 
8:17 AM
Morning data shovellers
 
Morning
 
Masters of the DB-verse
 
8:39 AM
Data travelers
2
 
9:36 AM
@McNets Yes they can be merged. It's probably the most common case.
150
Q: How can one link / merge / combine / associate two accounts / users? (Anonymous / unregistered / cookie, or OpenID / registered)

pauldunlopI accidentally posted a question without being logged in. Is there a way to associate that question with my account? Return to FAQ Index For more information, visit "I accidentally created two accounts; how do I merge them?" in the Help Center. a.k.a. How can you link a registered accoun...

 
10:09 AM
Morning, Data Engineers
 
 
1 hour later…
11:18 AM
@PaulWhite ok, thanks
 
 
1 hour later…
12:33 PM
anyone know off hand what the "Red Dot" on the review queue is? does that only pop up when the backlog starts to get out of hand?
 
85
Q: When I look at the review icon, I see red

TylerHI like the new notification version of the review icon, which I'm guessing is still undergoing A/B testing: It does a better job at saying "attention here, please" than the old icon which was almost indiscernible, and certainly easily ignored. What it doesn't do is tell me any useful informat...

 
12:48 PM
herp. derp.
 
1:09 PM
@PeterVandivier It lags though
sometimes you still see a red dot and when you click it the queue is empty
 
eventual consistency, we meet again, at last...
 
1:26 PM
we meet at least
3
 
1
Q: Unique filtered index and surprising estimation

codersWhen I view the estimated execution plan for this query... SELECT * FROM tabela2 x LEFT JOIN dbo.tabela1 t with (nolock) on t.Kod = x.Kod ... the Estimated Number of Rows property for [table1] shows 10. Why is this? In the schema (copied below) I have a unique index AKF_tabela1_Kod. Sho...

Could it be caused due: ALTER DATABASE test SET AUTO_CREATE_STATISTICS OFF;?
 
2:27 PM
That's only for statistics that don't yet exist (i.e. columns without indexes) --> _WA_Sys_12345678_12345678 statistics.
A good example is the Person.Person table in the AdventureWorks2012 database.
 
@hot2use As soon as you set ALTER DATABASE test SET AUTO_CREATE_STATISTICS ON Estimated number of rows change to 1
 
Does executing the query result in any _WA_Sys_12345678_12345678_ statistics being created in the tabela1 or tabela2?
An AUTO_UPDATE_STATISTICS OFF would result in the statistics being stale, otherwise I'd have thought that the statistics would be up-to-date. Unless of course AUTO_UPDATE_STATISTICS_ASYNC has been turned ON.
Seeing as tabela2 doesn't have any indexes and AUTO_CREATE_STATISTICS is OFF then the Query Optimizer wouldn't have anything to rely on. But the predicate in tabela1 should have valid statistics.
Bug?
 
2:44 PM
"Because the automatic statistics creation option has been disabled on the database. If there is a dedicated index filtered (seek) on a column , why do I need the statistics not filtered on the entire column? Rhetorical is the question :). The question is, why SQL Server does not correctly estimate the number of rows based on a unique filtered index? Is this a paradox that uses a unique index and estimates a value greater than 1?". So the user turned auto-create off on purpose
 
`object_id name stats_id auto_created user_created no_recompute has_filter filter_definition is_temporary
----------- --------------------------- ----------- ------------ ------------ ------------ ---------- ------------------- ------------
224719853 _WA_Sys_00000003_0D64F3ED 4 1 0 0 0 NULL 0
272720024 _WA_Sys_00000001_10416098 2 1 0 0 0 NULL 0
Can I use mono-spaced text here?
 
YEs. CTRL + K when entering text
 
object_id name stats_id auto_created user_created no_recompute has_filter filter_definition is_temporary
----------- --------------------------- ----------- ------------ ------------ ------------ ---------- ------------------- ------------
224719853 _WA_Sys_00000003_0D64F3ED 4 1 0 0 0 NULL 0
272720024 _WA_Sys_00000001_10416098 2 1 0 0 0 NULL 0
 
Enter Text then hit CTRL + K
 
object_id   name                        stats_id    auto_created user_created no_recompute has_filter filter_definition   is_temporary
----------- --------------------------- ----------- ------------ ------------ ------------ ---------- ------------------- ------------
224719853   _WA_Sys_00000003_0D64F3ED   4           1            0            0            0          NULL                0
272720024   _WA_Sys_00000001_10416098   2           1            0            0            0          NULL                0
 
2:49 PM
What columns are referenced in the statistics?
I guess his non-clustered index is the cause: (Kod) INCLUDE (id,kol1) WHERE KOD IS NOT NULL. The index isn't being used and because the query doesn't have an index/statistics to go with, then the QE is unable to estimate the number of rows.
He could have basically omitted the index and would probably have the same results (whilst AUTO_CREATE_STATISTICS is OFF).
 
@hot2use a user has answered pretty much the same thing
 
object_id   name    name                      auto_created st_col_id   col_id
----------- ------- ------------------------- ------------ ----------- -----------
224719853   tabela1 _WA_Sys_00000003_0D64F3ED 1            1           2
224719853   tabela1 _WA_Sys_00000003_0D64F3ED 1            1           1
224719853   tabela1 _WA_Sys_00000003_0D64F3ED 1            1           3
224719853   tabela1 _WA_Sys_00000003_0D64F3ED 1            1           3
272720024   tabela2 _WA_Sys_00000001_10416098 1            1           1
 
3:10 PM
Seems I'm a pretty big deal, check out the meta leaderboard
 
@hot2use thanks
 
@billinkc s/4/2?
 
Yes, but that's the link SE sent this morning
 
curious & curiouser
i feel like at some point, it might be worth the trouble of considering a popup saying "have you considered PasteThePlan.com" for new users with sql-server tagged questions
 
3:31 PM
0
Q: Making a button disappear after clicked in php

Joshua Brischkemy current code is this button type="submit" class="btn btn-primary btn-lg" name="submitbutton" value="Submit" onclick="this.style.display='none';"> Execute Procedure </button I am trying to make the button disappear after it is clicked and that onclick even works accept when the page upd...

^^^^ SO question
 
@billinkc I got a mail too. But when I click on the link, I get the 404 message (panda).
So is meta ranking only for mods?
 
Only Meta StackExchange appears in my list.
 
:¬|
Going home. Had enough.
 
3:57 PM
@McNets tsk tsk
 
@TomV not a valid link?
 
@McNets I noticed your comment after posting an answer
 
@TomV oh, no problem, I've deleted it.
 
 
1 hour later…
 
1 hour later…
6:33 PM
In the SQL Server change tracking feature, how can a part of your system actually get a notification that a change has occurred?
I'm reading the docs and it says: "Typically, to determine data changes, application developers must implement a custom tracking method in their applications by using a combination of triggers, timestamp columns, and additional tables. Creating these applications usually involves a lot of work to implement, leads to schema updates, and often carries a high performance overhead."
Specifically, I am looking to evaluate using this feature to start some ETL to pull changed data from this OLTP database to a separate analytics database with as few interventions in this database as possible (since it's live clinical medical data). But with a roll-your-own system which includes triggers, there is obviously a TRIGGER - so you know when a change event has happened. Do you basically have to poll the CHANGETABLE() function?
I also am evaluating CDC, but that has a lot more objects it creates and it seems to require SQL 2016 before it's available in the Std edition and we have some customers on SQL 2014 Std.
 
7:00 PM
@CadeRoux basically yes
It has restrictions on schema changed etc
But you could base your etl on the change table output
It's a bit of a broad subject
 
7:45 PM
So I basically just poll CHANGETABLE from my analytics database and pull data in. Out of the 3 systems, CT, CDC or Temporal tables, Temporal tables are out (for now) because it requires re-engineering of source system. CDC is out because of version/edition issues and also the level of schema changes and CT is the lightest weight of them all and requires the least changes to the schema of the source system.
When they talked about CT being synchronous, they are just talking about the commits to the change tracking? There's no other more significant process there, so the advantages of CDC are that all the other additional features it has like storing every change are queued and there could be a slight delay, and it actually needs to be asynchronous to reduce load, but that the CT synchronous load is still very lightweight, right?
 
8:08 PM
@CadeRoux I'm not familiar with the implementation details but dynamics uses cdc for the retail clients to only update the remote stores with changes and from the looks of it I think you are right
@McNets what I meant is that the time it took to Google that link was only a minute less than the time it would take to write something in the answer box
 
@TomV ;-) I know, I'm sorry.
 

« first day (2898 days earlier)      last day (1955 days later) »