Sql Record number
in [Programming SQL]
|
From: Tony Rogerson on 20 Oct 2009 16:48 The original op question was around data mining to which you stated "SQL Server is not in that game; you ought be using Teradata, KX, WX2, etc." I asked you to qualify your experience within SQL Server that qualifies you to rule this product out and berate it - you replied SQL Server doesn't do prediction (which of course now you have actually done the research you know now is wrong) - finally after several attempts you replied that you haven't any. I am not a statistician, nor have I claimed to be one; I have 5 years DB2 on the mainframe, 3 months Oracle and 16 years SQL Server experience - all full time, all fully trained from the best and all kept current - I work around 60 hours a week in the product space AND the database sector servicing my clients and keeping current by researching. SQL Server is a product set, what you meant to say was its not built into the relational engine - which is your problem, you are blind and see only SQL and see only SQL Server as a departmental store and retrieve database - which it isn't and we've told countless numbers of times with references etc.... The rest of your post is just drivel from an old man who's had his day and is upset that he cannot keep up with the industry and nobody will employ him except for the odd training gig a handful of occasions in a year. Who's going to be more current? Somebody like myself who is in the industry full time, spending on average 60 hours a weekend on clients, research and other stuff around the product and the field, or - somebody like yourself who thinks he knows it all already, rarely works and is so out of touch its laughable. But - take my word --celko-- I'm here to stay - for every post you get wrong or berate the op I'll be there berating you and questioning you - making absolutely sure other folk see you for what you are. --ROGGIE-- "--CELKO--" <jcelko212(a)earthlink.net> wrote in message news:30c9962e-b614-4553-9ce5-db102be45748(a)k26g2000vbp.googlegroups.com... >>> Precisely - you've never even installed it; and judging its took you a >>> few days to reply you've had to actually go and research what Data >>> Mining comes built into the product. << > > I don't have a Data Mining client on SQL Server. But I did my > research on MS stuff and will still go for SAS or SPSS when I do > serious work. > > The Data Mining is not built-in to SQL Server; it is stuck on the > sides. Are you familiar with Floating Point Rounding errors? I don't > mean that you know they exists; do you now about the **corrections** > that should be applied ? The easiest is for addition (sort on > absolute value, sort from low to high and add in sequence) , then they > get complicated -- logs, square roots, etc, require different > corrections. > > This is not so bad when the data is less than a million rows. But I > work with more data than that. > >>> Yet another sign of your arrogance in believing that you can just go on >>> in life without researching and keeping current. << > > LOL! One insane rant you call me an academic who is all research and > the next rant I do no research. I get called in by start-ups so that > I am ahead of current. > >>> And given you do all your work on Oracle, DB2 and Teradata - why aren't >>> you plaguing their forums with your frequently incorrect drivel - oh, >>> because you don't work that often aside from the odd training gig! << > > Tony, I counted 17 different SQL products that I worked on or > consulted for the product developers over the years. I am the "SQL > Guy" and not the "one-product-only Guy" instead. > > Oh, before I get trapped in the "Troll Attack" that you are doing. I > should know better than to reply to someone who rants about > Heliocentric models of the Universe in a DB forum. But let me ask: > > 1) Would you mind answering my previous questions about YOUR > credentials for data mining? Do not go into details -- just the title > of your last book or article, or the last project (year and enterprise > area). > > 2) Where did you post your query for Suduko query? You siad it woudl > take you 5 minutes to write a better answer. > > For those "Troll Trackers" coming in late, a friend of mine (Richard > Romley) wrote a single query stored procedure that generates ALL valid > solutions. It was demonstration that an 81 parameter procedure call > was not hurt by a long parameter list (another myth among newbie SQL > programmers), that SQL Server can do an 81 way self-join, that simple > direct search conditions are faster than doing bit manipulation, etc. > > It showed a lot of the strengths of SQL Server. > > And that a lot of Sudoku puzzles that are published have multiple > solutions. ARRGH! Tony's reply to Richard was to ask for a solution > to a grid that had zero to nine across the top row. Think about it; > This is a request for an isomorphism of all valid grids. The number > of possible 9 by 9 Sudoku grids is N=6,670,903,752,021,072,936,960 > which is approximately 6.671�1021. > > Gee, I wonder why Richard told your demand was impossible in the real > world? Duh? Perhaps because BIGINT tops off at > 9,223,372,036,854,775,807 is a consideration. > > The nature of the insane or a newsgroup troll is they only throw out > insults and misquote facts that are available. They never answer > questions ask of them. They spit bile without explanation as to why > there was an error in the poster's statement. > > NOW > > 1) Post your Sudoko query/SP > > 2) Tell us, AGAIN, how many years have you been employed as a full- > time statistical analyst? How many credit hours in stats do you have > in your Master in Math? How many books on data mining have you > written? Any articles-- academic or popular press > > > > > |