No.
TRUNCATE and DROP are almost identical in behavior and speed, and everything else proposed here as an answer is incorrect.
I am astonished at the number of falsehoods on this thread, but then again when I first read it some months ago I did upvote some of the answers myself. Why? Because ...
and read the comments on the now-downvoted (after my expose) answers to that question
I was originally going to phrase it "if you'd spent ten second looking at the page, the INSTEAD OF UPDATE link should've jumped out at you", but thought I probably shouldn't be snarky
@JNK "I try to avoid them since they can be a maintenance issue (i.e. index gets updated but isn't present on the actual table so may not get disabled when needed)" << I don't understand, help?
@SQLkiwi - I've had issues before with slow inserts in a large table because of indexed views that someone else created. We have as part of our process programatically disabling NC indexes for inserts (monthly loads of 50-100m records for ETL). They had built some indexed views to solve some small corner case and didn't account for the ETL
So I'm disabling indexes on the table but not the indexed view which references the table and still gets updated when the new records are loaded
@JNK Thanks for clarifying - I didn't want to get into a discussion about what you meant on the question. It's not much more work to programmatically disable indexes on views too, but I do see what you are saying.
@JNK Yep, everything is a compromise. If that COUNT(*) query runs many times per second, an indexed view might be a big win. In other scenarios (e.g. very heavy updates, infrequent selects) it would not.
@jcolebrand @JackDouglas Did we establish a guideline for the product/product-version tags e.g. sql-server & sql-server-2008 both being included on a post or just sql-server-2008 suffices? I see there are 5 suggested edits which are just adding the sql-server tag
The app I started off deploying was ported to Oracle in the late 90s because we were having major issues with Sybase - lock promotion was killing us on our busy huge tables. Kinda glad. <3 Oracle
@MarkStoreySmith I just accepted all those edits before checking in here - we probably do need a policy (lets take it to meta!) but one good reason for double tagging is the 'top users' list for each tag
I've noticed a couple of questions that have all of the tags:
oracle
oracle-11g
oracle-11g-r2
Ideally, the tagging would support hierarchies, and know that 'oracle' is a broader term of 'oracle-11g-r2', so we didn't have to put on the redundant tags, but I don't think it currently does.
So, ...
@NickChammas Yes. Some questions will not be specific to individual versions, and having generic tags will help searches, even when the answer is specific. If someone looks for 'Oracle' but there is no 'Oracle' tag on the question then it might get missed in search results.
Can you set up a synonym so a general tag such as 'oracle' is equivalent to all the specific version tags but the equivalence doesn't go the other way?
How about with a developer question after the main introductory blurb:
@JackDouglas - do you mean something like this?
Database Administrators - Stack Exchange is for database professionals needing expert answers on all advanced database-related topics concerning either traditional SQL RDBMS or...
Hi,
we already have tag synonyms, what about having also "single-directional synonyms"?
For example: We have tags like C# or WPF. Both of these subjects are inevitably related to .net and should be tagged as such. But many people don't respect this and often tag C# related question only with C#...
@jcolebrand if we leave sql-server-2008 some on r2 will use it without thinking
they may not notice the translation
@ConcernedOfTunbridgeWells that is better - but my main thought is that we should leave that bit out and address it after the BI merge. I can't think properly about two things at the same time :-)
1) We'll entertain site name changes...so long as it's not just slapping on BI onto the current title, haha. 2) Adding the mod could probably wait until the next election, but we're not opposed to adding one pro-tem. Mostly, Word of God says that y'all are probably overthinking it :P 3) Let us know when you make it Officialness and we'll email all the people who followed/committed to the BI proposal. :D
So what we need first before we update the FAQ is some meta Q to point to, saying that this is acceptable, this is not, etc. Example posts, effectively.
> I'd suggest just doing the soft merge, then, and worry about the naming issue later. They should start with a featured meta post announcement and the appropriate FAQ changes. They'll want to make sure the FAQ doesn't "over sell" the change. FAQ authors have a tendency to do that: make the most reacent additions into the most prominent part of the site
@jcolebrand Well technically even that can wait until after the faq changes and the merge - we just need the meta-meta question to be posted so we can link to it.
@jcolebrand I think so - @ConcernedOfTunbridgeWells can you live with that?
@JackDouglas I don't mind. The B.I. stuff is included in the draft FAQs and IIRC isn't any different between the two drafts anyway. If you want to deal with the 'development questions' issue separately I don't think it makes much difference.
I just love love love it when people reference the SQL92 standard without knowing what the fuck they're talking about. It's right here, fools! You can read it whenever you like!
@gbn Read the comments on this question. In this case I wouldn't call it fuck-wittery, but more so hand-wavey implications that are not supported by the SQL92 standard.
Viewing deleted comments was implemented, now what about being able to restore them?
As far as I'm concerned, make it a simple button on the same form where we can see the deleted ones.
that's just a sign of lower self-esteem, which would coordinate with my failure for roughly a week to go and purchase my androgel (mostly because of insurance regulations here in the states combined with a recent shortage at that particular pharmacy)