@AaronBertrand Yes, noticeable increase in the dross. It was ~2.5 years for me to hit 1000 close votes. The next 500 came in 5 months, during which I was much less active.
I think we're just about keeping on top of it though
To get the result that you want, it appears that you'll need to use a FULL JOIN. A FULL JOIN will join your 2 tables on the bln column but it will return data from both tables, even if the value of the column doesn't appear in both tables.
The basic syntax of the query will be:
select
bln = ...
How can i access / view the table names present in a database.Currently I am restricted on privileges(Deny permissions on View Definition ) to view the tables.
I have to refer to a table containing the column but unable to do without knowing the table name.
(PLS:00201 'Identifier must be decl...
There is no concept of a Master/Slave in Vertica. It seems that you are after a DR solution which would give you a standby instance if your primary goes down.
The standard practice with Vertica is to use a dual load solution which streams data into your primary and DR instances. The option you'...
hi, I have a sql server database in my server and i need its copy of it in my local machine, updated daily. How can I do it in a bandwidth efficient way?
Yes, we get your requirement, but if you are doing dev work on the local copy, it's not possible to use differential backups to update the local copy. Why do you need new production data every morning?
@rjv so you're going to restore the backup every morning? Ok, so you can use differential backups from the production server every morning, as long as you don't take any more full backups there.
Copy differentials from prod to local. Restore to differential to Db1. Write sql/ssis packages to identify changes and apply to Db2 which is where I assume you're working and modifying data, fixing etc. Keeps the restore chain thing going and gives you current-ish data
If you're pondering log shipping you need to go READ ABOUT IT. No, it does not ship the whole transaction log, it ships the log backups (.trn). However it is still unclear exactly what you are trying to accomplish. With log shipping you can't make any changes to the copy of the database, unless you stop log shipping - so what is the purpose of having it on your dev machine? Is this supposed to be disaster recovery? Because the approaches you seem to want do not allow you to do any dev.
@rjv you want disaster recovery to your local desktop over a network pipe that can't handle 10GB per night? Sorry to be blunt, but you're doing it wrong.
I built a package in SSIS that uses a script task to open an Excel file, format, and refresh some data in Excel. I would like to have Excel visible when the script task is running to see if Excel gets hung up which occurs all the time. Is this possible? I am converting a process that is calling E...
> STARKE, Fla. — A female passenger in a semi-truck was naked when it slammed into a school bus Monday afternoon, according to the Florida Highway Patrol. Seven students were taken to the hospital after the crash in Bradford County.
@AaronBertrand i've read conflicting reports ... one is where abdullah acknowledge that he got the penalty for sliding into the endzone. if he got a penalty for the prayer then that's a bad call and shouldnt have been called. if he gets fined then it's worse
@swasheck the announcers definitely said it was unsportsmanlike conduct because he went to his knees and bowed. Sliding? Really? People do all kinds of crazy crap when they score a touchdown, dollars to donuts if he didn't do the praying thing he wouldn't have had a penalty.
Of course we'll never know, now, because officials will start calling every single player who even thinks about touching the ground to make this call look better.
@AaronBertrand i didnt see the game so i dont know exactly how it went down but based on what i've read, it shouldnt have been a flag.
@AaronBertrand they're just going to start enforcing the no celebration rule more stringently. "see, this is why you can't celebrate. there's too much room for misinterpretation and hurt feelings."
@swasheck agree 100%, it was silly. Like Goodell, the officials are focusing on unimportant things and turning a blind eye to things that matter (like the offside and the late hit)
@swasheck yeah the slam dunk on the uprights too, WTF? Next you'll get a penalty for smiling after a touchdown because it will come off as gloating
@AaronBertrand wagging the dog. but hey - it's been bad all year. we had a 1st down against seattle taken away because their dt was in the backfield before the snap. that was a momentum-killer on that drive. the nfl is becoming more and more a farce.
Crazy, just heard Seth MacFarlane with some symphony orchestra singing "Singing in the rain" It sounded like a professional and until the announcer went on to clarify that yes, it's the "Family Guy" Seth I was thinking what are the odds
We have an SSIS package that has a script task that does SFTP to another company through WinSCP. Right now the password is stored in the package as a variable in plain text. Obviously this is not ideal but I'm trying to get a feel for what the best route is for going forward. Using SQL server encryption and loading the variable at runtime or maybe storing it an encrypted Config file. Thoughts?
k. The SSISDB catalog makes the storing and management of sensitive data much easier as it's more native than the manual steps you'd use with pre 2012/package deployment model
Best practice, can't say but we used native SQL Server configuration. Extended the base table to include a column Senstive. The DBAs then denied read access to the developer group for any rows that were marked as sensitive. The service account could access those rows so things worked on servers.
For development purposes, we all ended up knowing the password. Less than ideal, I know. SFTP seems to be a finicky thing from my most recent experience with it.
I like that MS thought that preventing the next ILoveYou outbreak by denying access to PS execution was a good thing but in practice, it's a punch to the junk
@MikeFal yeah they made a home brew SSIS management application called PANDA some 10 years ago. It's a solid effort and does some cool things. However I feel that it gives more limitations than it actually helps.
@swasheck I think it has to do wth performing aggregations separately earlier in a workflow
I.e. if I need to get
SELECT A, MAX(c), MAX(d) FROM TblA JOIN TablB on somekey JOIN TablC on SomeotherKey
group by a obviously
If say D is functionally independent of C or A, and you just want to get a value from that table matching the join key, you can do the aggregate first then join to the intermediate result set
Although it looks like I won't be able to get the data needed, this is the best answer and, therefore, I will mark it as such. — Wes Crockett13 mins ago
@Brandon Unless you specify a nolock hint, there will always be a lock. <-- this implies if you use nolock there might not be locks. I was merely clarifying. — Aaron Bertrand ♦46 secs ago
Mine is doing hard work for the optimizer before joins
If you know you need to get the sum of some measure for each day in a month, it may be more efficient to group/sum it all first then do your other filtering depending on indexes etc
@Shanky Mark is correct. The KB article is not brilliantly worded. It is referring to some types of locks that are held to end of statement or end of transaction (e.g. an exclusive lock protecting modified rows). In general, locks are released eagerly, not in the way described in your answer. Sorry.
@PaulWhite i havent seen them --- just saw the headline that said that microsoft was skipping the windows 9 name because this version was just "too big"
But if I've ever access a MS property, then yeah I get redirected through the login thing and Oh, It's expired, log in again. WOULD YOU LIKE TO TAKE A SURVEY?
@Paul Ok well thank you I agree to Mark's statement he is correct I read about that but OP question was can locks be held and I guess so it does or I am not correct ?
@Shanky Arguably, in certain cases. The way I read your answer you seem to be making a different argument. Perhaps you can reword it a bit to make it clearer?
@billinkc Yes, maybe it has something to do with whether I logged in previously but now the cookie has expired, or something. I guess I'll look next time
i'm trying to think of a nice response to this email i just got
> Aight.. so I did it as a sub query, the problem is when I move the where clause to outside the query it takes forever to run… inside the subquery it works fine. But the problem is with it inside I can’t run generically for any (parameter combination)
sigh In the time it took you to ask this question, you could have checked both
Should I look at the system clock of the CLIENT: 192.168.45.242 or the server it is trying to connect? The original image is for the server it is trying to connect to. — Debbie31 mins ago
"I think something changed at the o/s level" You think you did or you did. These are the people that think computers are magical machines and hitting the sides will free up whatever's gummed up the works
Would any of you disagree that inserting 600 MB with a single transaction during peak hours into a table that has lots of concurrent reads is probably a bad idea with OLTP, absent further details?
I assume the insert will escalate to a table lock and will either be waiting forever to get lock or everyone else will be waiting while data gets inserted
But maybe that 600MB is what must be done right then and there as people are making operational decisions based on it. "We're out of chicken nuggets! ORDER MOAR (but if you'd let me get my BA query run, you'll see we're flush with nuggets)"
Anyhow, the reason I asked is that someone was upset that someone inserting into an invoices tables during the day caused blocking because "SQL Server should be able to handle this." My response was to consider RCSI or batch it up. I thought about suggesting a lock hint, but was worried that would just make it worse.
Service Broker is an interesting idea. This is a vendor app so I'm not sure how difficult it would be to fit in.
RCSI/SI would be the most general solution, though it depends on the priorities. I worked on a very busy OLTP system once that only ever performed changes (of any size) as single row operations with a small WAITFOR delay between each one.