I've looked all around the internet and I can't find a working solution? I know there's no IF in SQLCE, so I'm trying other different stuff, but no luck. Does anybody know a solution that does work?
Thanks!
I voted to close because it looks pretty close to two other questions (which I referenced in my answer). But I realized that the OP doesn't want to put the logic in his app code (which is what the other questions were asking). I think he will have to put it in his app code regardless, but in retrospect I don't think the vote to close was fair.
yeah I made a joke about what SQL CE supports and what it doesn't. I really think it is a misnomer having "SQL Server" in the name. Sure some of the SQL dialect is similar but it really is a different product.
This is general discussion - questions about the site are ok of course but this is not a place to ask your database question. Also you shouldn't submit questions here unless it goes unanswered for a long time or you have an issue or concern about the question, answers, comments, that kind of thing. At least that's how I understand it.
Well I guess where questions about the site should go is not really that well-defined. For example, I've given up on meta as an approach to change functionality, so I'm just as likely to ask a question here (like why can't we undo a close vote?) then pose a new question on meta. YMMV.
I'm going to get answered much more promptly, I'm not going to get down-voted by all the people who think my question is dumb or that the request shouldn't be implemented, and I can carry on a conversation about it without getting chastised. :-)
When people don't want to talk about it anymore, they'll tell me to shut up.
Yeah. I get that. I was just wondering about "emergency" questions and such. I seem to have solved my own problem, but was afraid to do it ("disable" replication by removing articles from a subscription and then "enable" it by adding the articles back).
So the point of chat is just to rub elbows with some like minds?
Yep. If you want immediate help on SQL Server-related things, I suggest submitting a question and asking for help about it on twitter using the #sqlhelp hash tag and pointing folks at your link. This will maximize your potential audience, since not everyone here is constantly hitting refresh on the "newest questions" list 24x7, but twitter is pretty heavily monitored in all time zones.
For general user population who are tweet-savvy (a small % of a relatively small number), who aren't asking 15 questions a day, I don't think it's a big deal. They are pretty funny about what they complain about and what they don't (e.g. promoted tweets are the devil but off-topic questions about active directory are ok).
I don't mean every five minutes post to the tag saying "hey, did you look at my question yet?"
I answer on #sqlhelp from time to time and as an answered on that resource I welcome requests that point to a well-constructed question rather than the gibberish that typically results from cramming a question into 140 characters.
as an answere*r*
Not that being on dba.se or SO means it is not gibberish, of course. But it isn't forced that way due to size constraints.
Well again, I wasn't suggesting that every single question should be accompanied with a post to #sqlhelp saying go look at my question. If you want immediate help on SQL Server-related things which should be a very small percentage of the daily questions (the subset of SQL Server questions asked by tweet-savvy folks who want immediate help and saw my suggestion here).
I'm not going to go post my suggestion on the FAQ. Swasheck stated it was a bit of emergency and I was just suggesting a way to solve that with the best of both worlds.
You know, when I read stuff like this it makes me kind of mad about the whole big hullaballoo that occurred the other night when I asked a question I knew the answer to and waited before posting that answer...
That link has meta.dba.se giving voted-up opinions about waiting before posting your own answer. Then, I see posts like this one about it being okay to ask for a specific answer to your question and I'm left kind of shaking my head...
@JNK Maybe just a few every day, and use your SO privs to tidy up the stuff before you move it. jcolebrand had some cogent stuff to say on that yesterday.
So, when SO looks at any other site, they have to judge if they're going to be unfairly dictating direction on another site, especially with all the mods and users they have, they can't just chuck everything databasey over here (or programmer related to programmers)
From there.
There are a few specific cleaning up issues that cause a problem to the mods.
I think a steady flow - I have a few more in my answers that could probably come across, but a trawl through various tags like 'sql-server', 'mysql', 'oracle' etc. would probably come up with quite a bit of good material.
We have 3 copies of the exact same database. All 3 databases have a Users table, and a User will always exist in all 3 databases with the exact same settings. Anytime we want to add or edit a User we have to update 3 databases.
Would it be a better idea to delete the Users table from databases ...
Alright, I posted on programmers since I was thinking it was related to software development, but thought afterwards it would be better on dba. Should I delete and re-post, or just vote to migrate?
I got that error when I tried to execute this query:
select *
from lc_history2 lc
where lc.view_name <> 'LC Main'
and lc.modifiedtime = (select max(modifiedtime) from lc_history2 l where lc.lc = l.lc)
and lc.modifiedtime >= sysdate-30
@JNK true, I have one of those too, but I don't have any key lookups on the other table either, and it's not clustered and it's not clustered, should be a covering index though
Just trying to understand what it means for my query plans
(By request, RE: earlier tag wiki discussion, anyone can suggest a tag wiki, but only > 20k can edit without suggested edit approval; tag synonyms on the other hand do require score > 5 in the given tag to be voted on, irrespective of rep)
Oh, on one table I'm using the clustered index scan (which has a crazy high cost) and on the other I'm using an index seek (which I'm pretty sure is covering) and the index seek is way faster despite similar # of rows and columns.
@TimStone thanks for the clarification. I was wondering why I could approve edits but not make edits! I missed the part that I can make edits but they still have to be approved!
Well there's my problem, the query was using a generated index where the "included columns" were the key fields and the random associated columns were the "key". So that's interesting
Interesting, SSMS insists the index it suggested is "missing" but I get better performance with a different index. I'm betting if both existed it would use the slower index it suggests?
It's hard to ignore the SLEEP_TASK wait type when it shows up as the last_wait_type in sys.dm_exec_requests for a poorly performing query.
It's got to mean something in that case, question is "what?".
@MarkStoreySmith FWIW I've worked for clients in insurance, public sector healthcare, courts system (sometimes referred to as CJIT here), property management, building supplies and various other public sector work.
@gbn I don't think @MarkStoreySmith's client has heard my rates yet either, although they're not anything special by B.I. standards.
@JNK without opening the kimono too much, B.I. work in Insurance in London tends to go around the £400-550/day range, depending on the client and the role.
Banking goes a bit higher than that. More if you're into doing quantitative analysis. Project management often goes around £500-750/day for B.I. projects.
@ConcernedOfTunbridgeWells How did you find public sector clients? I was constantly torn between the "they're paying me, good" versus "its my tax they're wasting, aargh"
If you're under 32 (I think) you might be able to get a working holiday visa and go contracting for a couple of years. They're a bit hard on immigration these days so skilled migrant visas aren't so easy to get now.
@JNK Rent the house out. I made the mistake of selling mine before I came here. I don't think I should in hindsight. Compared to London prices carrying the mortgage would have been peanuts.
Well, DBAs are still in demand in London, and data analyst/report development roles are pretty common. They can even pay quite well on the contract market if you've got industry experience.
Is it you that's working for the health insurance outfit?
@JNK Bone up on the terminology and the industry and you might be able to break the chicken-and-egg problem getting into the industry in London.
A data analyst role can pull £350-400 a day on the contracting market in insurance in London. That might be out of date though, and it doesn't really help you much with a pregnant wife in tow.
The problem is getting your foot in the door. Once you've done that you're in. You've also got a fair slug of rep on Stackoverflow, so you can probably spin that a bit as well.
I've looked all around the internet and I can't find a working solution? I know there's no IF in SQLCE, so I'm trying other different stuff, but no luck. Does anybody know a solution that does work?
Thanks!
We have 3 copies of the exact same database. All 3 databases have a Users table, and a User will always exist in all 3 databases with the exact same settings. Anytime we want to add or edit a User we have to update 3 databases.
Would it be a better idea to delete the Users table from databases ...
I got that error when I tried to execute this query:
select *
from lc_history2 lc
where lc.view_name <> 'LC Main'
and lc.modifiedtime = (select max(modifiedtime) from lc_history2 l where lc.lc = l.lc)
and lc.modifiedtime >= sysdate-30
It's hard to ignore the SLEEP_TASK wait type when it shows up as the last_wait_type in sys.dm_exec_requests for a poorly performing query.
It's got to mean something in that case, question is "what?".
