Databases do not return rows in a given order unless you supply an ORDER BY clause in your query. The order of a SELECT * FROM MYTABLE; query is undefined.
Apologies for the simple answer!
Too mean?
Sure one of you SQL Server deviants will explain how data pages/caches work etc
I have a good story about my previous answer in here
Imagine you have a telco that has a prepay system that has a batch that removes unspent free minutes/SMSes at midnight each night, then gives them their months allocation. This is driven by a config table. A maintenance at 3am one night deleted the rows and re-inserted them.
The next night I got a call at about 1am (I was on call) saying there was a problem
I looked an nobody had any free minutes
The rows had been re-inserted and happened to end up in different blocks
It'd been just lucky that it'd worked the whole time, as the proc that selected had no ORDER BY clause.
Everyone had their free minutes/SMSes added, then confiscated :D
What a clusterfuck to clear up. Christ
Had to write a load of mental code to re-credit people that'd been charged (out of their prepay balance) and take it out of the relevant SMS/data/voice balance
That wasn't a fun few days
My other favourite was a muppet that shouldn't have been able to touch a Unix box in the first place typoed "hostname - a" and ended up changing the hostname of the box to "-". Queue ORA-600 errors all over the place and lots of angry call center calls (from landlines, as he'd stopped call authorisation for prepay working). I found that problem quite quickly!
@Phil the community did fine with that one actually.
@Phil StackExchange is a web culture. The site here may be focused on dba's, but overall our audience is still restricted by and large to those writing on the LAMP or ASP.NET stacks, sadly. We have very little SAP style users, apparently.
So let me get this straight - when a bunch of people tell you that many experts consider a row and record to be different concepts, you crap all over it and refuse to acknowledge it because it didn't come from an official source (even though one such quote came from an MS Press book written by the one and only Itzik Ben-Gan). Yet your evidence here consists of two excerpts from random junk on tech target? Double standard much? — Aaron Bertrand2 hours ago
"who are arguably experts in the field"
he didn't say that, but Aaron probably should have ...
could anybody suggest we which data base (between big data and Oracle database system) is best for analytic mining of telecommunication data (user profile and Their call details)
@Pattinson Only if you can tell me what the best bucket to bail out the ocean with is.
Seriously, your request doesn't make a lot of sense.
First off, who is this for? (I don't need a name) Is it a regional carrier? Do they have more or less than a million customers?
Will they need to do itemized chunked data analysis, or will the system merely have a few fields per record, such as "time of call, destination of call, origination of call, length of call"?
I mean, you could be recording second->second including nearest three towers that recv'd the signal, current timestamp within the call, origination timestamp of the call, dialing number, rcv'g number, conversation ID, ICCID (is that the name of the phone id?) of each party if it's mobile, quality characteristics for each tower, and effective data rate of the call.
Do you see the difference? One is trillions of rows of obscene amounts of data, one is hundreds of millions with relatively small data sizes
Assuming you need to do analytics on trillions of rows of complex denormalized data, I suggest the first step is to hire a professional DB Analyst or two and let them consult with what you need to do analytics on.
If you need to do analytics on hundreds of thousands of records, you're in the right spot.
@Pattinson does that help at all?
Because, to be honest, I'm not entirely sure what you need, since you half-asked one question in a quiet chatroom
@swasheck no need to worry, and sorry if I seemed to take sides against you. Read this if you haven't already and you'll understand where I was coming from :)
@JackDouglas Of course the standards aren't freely available, and the early articles from the 70s and 80s aren't online, but I found what appears to be an early quote of Codd's rules which do mention records. However, recent work by Date definitely appears to use row, and of course for all the reasons I mentioned in my answer, it is generally a better idea to use row and that is the excepted usage by most (if not all experts), including Celko, who makes a big deal about it, as already noted.
I think especially with no sql paradigms and a variety of "databases" which take elements of network, hierarchical and object databases, I think it's best for relational databases to use table/row/column to further reinforce that the relational model is one which is largely standardized and consistent - in terms of nomenclature, capabilities and well-understood rules like normalization - certainly compared to the fruit salad of other solutions out there.
My .ldf file size keeps increasing. I already tried to shrink it but after that it just grows again.
How can I permanently limit the filesize or perhaps even disable it entirely (since I don't think I'll ever use it).
@JackDouglas I only hear it from non-technical database users involved in design - never from anyone talking about the database. Which is another thing - they don't know the physical design, so I always use row and column to mean physical row and column.
oh look, more rep for saying something generic ...
@CadeRoux woot!
We're appreciative of the community participating in events, even if we don't all manage to get behind them. You guys make the site. (I'm not really a dba, am I? I just kinda mangle code from time to time. Ergo, I get to play here, but I really ain't the target community)