May 18, 2023 at 3:23 am
I agree with Eirikur on all this... I believe that a whole lot of people resort to the likes of Python because they don't actually know how to use SQL Server/T-SQL.
Another reason is they believe in the total BS out there like "To a hammer, everything is a nail" without understanding that when you're trying to drive nails, you should use some sort of hammer. The other extremely tired saying of "Just because you can do something in SQL, doesn't mean you should". Pure Rubbish. It should be "Just because you can't even spell SQL, doesn't mean you shouldn't learn how". It is, after all, the language of structured data. Perhaps we should start the new mantra of "Just because you can do something in Python (Powershell, C#, VBA, flavor-of-the-day, etc) , doesn't mean you should".
--Jeff Moden
Change is inevitable... Change for the better is not.
May 18, 2023 at 3:27 am
Just a post so that the 1st post on this page actually shows up. I wish they'd fix that and also finally get the pages things fixed, which is simply a problem of posts being marked as SPAM not being counted for the page counts.
--Jeff Moden
Change is inevitable... Change for the better is not.
May 18, 2023 at 1:51 pm
... Perhaps we should start the new mantra of "Just because you can do something in Python (Powershell, C#, VBA, flavor-of-the-day, etc) , doesn't mean you should".
Love it, 100% agree, two thumbs up, you hit the nail on the head with that statement(Ha Ha). Exactly how I feel Jeff. 🙂
-------------------------------------------------------------
we travel not to escape life but for life not to escape us
Don't fear failure, fear regret.
May 18, 2023 at 3:51 pm
...
All I am trying to say with all of that, is be careful what you recommend to others as they might go down the wrong rabbit hole based on that advise and suggesting that someone use Python or R instead of the Sequel Server natural environment is definitely advising someone to go down the wrong rabbit hole.
I wouldn't advise anyone to use or not use Python. You can disgree with my points, but they're from your perspective and the problem space in which you're working. If you have an application and you are connecting to it to do data analysis across large volumes, then likely SQL works well. If you've set some things up, this is great.
If you do ad hoc analysis, or you have people that are semi-technical and want to often change the way they might look at data, especially with some visualizations, then I might not think that. If someone sends me a CSV of data from somewhere, I think of these things:
If I wanted to do something and show people, I might pick R or Python if I did this type of thing, not with this file, but this type of thing, often. I can quickly load into memory, I can manipulate data, I can create visuals, and I can drop my code in a notebook that others can use to do the same thing or see results/visuals.
In an ad hoc, lightweight way, I think R or Python are much better. In fact, I had to load the Summit sessions from a CSV at one point and I tried both SQL and R. R was much cleaner and quicker, even with my having to look up a bunch of syntax.
It's not one or the other, but asking people what they are trying to do before deciding on how. I assume that's what most of you do here. You don't just say "use SQL Server". Even Jeff, who beats on other tech, thinks about why. He has reasons why in many cases, using C# or Python or PoSh, isn't better than SQL.
May 18, 2023 at 3:55 pm
Just a post so that the 1st post on this page actually shows up. I wish they'd fix that and also finally get the pages things fixed, which is simply a problem of posts being marked as SPAM not being counted for the page counts.
me too. I have complained and haven't gotten anyone to figure this out. I also feel allergic to PHP and haven't wanted to try and decode that.
May 18, 2023 at 3:55 pm
Jeff Moden wrote:... Perhaps we should start the new mantra of "Just because you can do something in Python (Powershell, C#, VBA, flavor-of-the-day, etc) , doesn't mean you should".
Love it, 100% agree, two thumbs up, you hit the nail on the head with that statement(Ha Ha). Exactly how I feel Jeff. 🙂
The caveat to this is that if you don't know how to do it in SQL but do know Python/R/etc., then use that. However, if it doesn't scale, learn SQL.
May 18, 2023 at 4:20 pm
Dennis Jensen wrote:... All I am trying to say with all of that, is be careful what you recommend to others as they might go down the wrong rabbit hole based on that advise and suggesting that someone use Python or R instead of the Sequel Server natural environment is definitely advising someone to go down the wrong rabbit hole.
I wouldn't advise anyone to use or not use Python. You can disgree with my points, but they're from your perspective and the problem space in which you're working. If you have an application and you are connecting to it to do data analysis across large volumes, then likely SQL works well. If you've set some things up, this is great.
If you do ad hoc analysis, or you have people that are semi-technical and want to often change the way they might look at data, especially with some visualizations, then I might not think that. If someone sends me a CSV of data from somewhere, I think of these things:
- I need a SQL instance and database
- I have to use some clunky import method, most of which are somewhat poor in SQL Server tooling
- I then need to run a bunch of SQL statements to clean or reformat data
- I can run queries, but visuals are few
- I can use Power BI, but then that's another whole project beyond simple analysis
If I wanted to do something and show people, I might pick R or Python if I did this type of thing, not with this file, but this type of thing, often. I can quickly load into memory, I can manipulate data, I can create visuals, and I can drop my code in a notebook that others can use to do the same thing or see results/visuals.
In an ad hoc, lightweight way, I think R or Python are much better. In fact, I had to load the Summit sessions from a CSV at one point and I tried both SQL and R. R was much cleaner and quicker, even with my having to look up a bunch of syntax.
It's not one or the other, but asking people what they are trying to do before deciding on how. I assume that's what most of you do here. You don't just say "use SQL Server". Even Jeff, who beats on other tech, thinks about why. He has reasons why in many cases, using C# or Python or PoSh, isn't better than SQL.
To be sure, my point was directed at people that tout and readily broadcast in a seriously passive aggressive manner that "Just because you can to something in SQL, doesn't mean you should" as a defense for some of the things that they use other technologies for even though they're the ones that are missing a point.
As you say, I do consider the Brent Ozar mantra of asking "What problem are you trying to solve?", especially if I'm the one trying to solve a problem.
An extreme example of this is that I have, in the past (long time ago and the code is long gone, thank goodness), used T-SQL to do a little screen scraping for data. It was a huge PITA and not efficient and I would never recommend that as an alternative to using Python for such a thing. At least I knew. In high contrast, a good number of rabid Python supporters that I've recently run into regularly bash SQL for things that SQL is actually much better at. The problem is, they're doing the bashing without actually knowing. They're just repeating the same crap that other supposed "experts" have also incorrectly stated. The other problem is, some of them are doing such in very public places (like the question I recently posted on this thread about a statement someone made) and that's going to be absorbed by people new to the world of data. In other words, it perpetuates the very wrong idea that "SQL is just a place to store some data".
Of course, that same kind of "expert" testimony happens a lot in the world of SQL. One of my recent favorites is people demonstrating how to combine a DATE and TIME column using a conversion to DATETIME for both and adding them together. Not only is there a loss of resolution but there's some pretty nasty rounding that can "move" the data to the next hour, day, week, month, and even year.
--Jeff Moden
Change is inevitable... Change for the better is not.
May 18, 2023 at 4:24 pm
Jeff Moden wrote:Just a post so that the 1st post on this page actually shows up. I wish they'd fix that and also finally get the pages things fixed, which is simply a problem of posts being marked as SPAM not being counted for the page counts.
me too. I have complained and haven't gotten anyone to figure this out. I also feel allergic to PHP and haven't wanted to try and decode that.
Heh... considering the other subject at hand, I believe that PHP can call a stored procedure. 😀
Anyway, thanks for the feedback Steve. I appreciate it.
--Jeff Moden
Change is inevitable... Change for the better is not.
May 18, 2023 at 4:50 pm
I agree with Eirikur on all this... I believe that a whole lot of people resort to the likes of Python because they don't actually know how to use SQL Server/T-SQL.
Another reason is they believe in the total BS out there like "To a hammer, everything is a nail" without understanding that when you're trying to drive nails, you should use some sort of hammer. The other extremely tired saying of "Just because you can do something in SQL, doesn't mean you should". Pure Rubbish. It should be "Just because you can't even spell SQL, doesn't mean you shouldn't learn how". It is, after all, the language of structured data. Perhaps we should start the new mantra of "Just because you can do something in Python (Powershell, C#, VBA, flavor-of-the-day, etc) , doesn't mean you should".
Unfortunately, too many have more knowledge of how to use their tools than the nature of the task at hand.
😎
Having said that, we all need to learn, sometimes by failing. I wrote a zip/unzip procedure in T-SQL, a lesson learned, it worked but soon figured out that it was not the right approach, just as Jeff has so many times stated, "Just be you can" isn't a justification 😉
May 18, 2023 at 5:23 pm
Jeff Moden wrote:I agree with Eirikur on all this... I believe that a whole lot of people resort to the likes of Python because they don't actually know how to use SQL Server/T-SQL.
Another reason is they believe in the total BS out there like "To a hammer, everything is a nail" without understanding that when you're trying to drive nails, you should use some sort of hammer. The other extremely tired saying of "Just because you can do something in SQL, doesn't mean you should". Pure Rubbish. It should be "Just because you can't even spell SQL, doesn't mean you shouldn't learn how". It is, after all, the language of structured data. Perhaps we should start the new mantra of "Just because you can do something in Python (Powershell, C#, VBA, flavor-of-the-day, etc) , doesn't mean you should".
Unfortunately, too many have more knowledge of how to use their tools than the nature of the task at hand. 😎 Having said that, we all need to learn, sometimes by failing. I wrote a zip/unzip procedure in T-SQL, a lesson learned, it worked but soon figured out that it was not the right approach, just as Jeff has so many times stated, "Just be you can" isn't a justification 😉
Too funny... that's one of the things I actually do use SQL Server for... makes logging of status and much more, real easy. So, in that case, I'll have to resort to what we all know... "It Depends". 😀
--Jeff Moden
Change is inevitable... Change for the better is not.
May 18, 2023 at 6:06 pm
Check out this video. The speaker gives a whole laundry list of reasons, a litany of epiphanies, on why SQL Server is not well suited for transaction processing. Instead, the better way, as he's telling it, is to create "future transactions" (jumbo shrimp?) to undo the cumulative totals he's running to facilitate calculating a discount for a promotion. Still using SQL Server afaik. At around the 5:50 second mark he basically confesses to not knowing anything about SQL Server. At around the 7:00 mark he goes to a SQL DBA for advice and he receives a stored procedure which works and performs well. Then he gets requests for more work and again we're told SQL Server is not well suited for exactly what he's doing. He feared his favored code was losing relevance and he laid awake at night worrying about stored procedures. So he invented this future events driven thing. Wow
https://www.youtube.com/watch?v=9ix0vVL34-M&t=2691s
Aus dem Paradies, das Cantor uns geschaffen, soll uns niemand vertreiben können
May 18, 2023 at 6:19 pm
Yes, why would anyone want to use that complicated futuristic hover car (aka T-SQL in this case) when we can spend 10 times the amount of resources to build a better dinosaur. I mean, then we will still have the dinosaur and its simplicity. Yeah it is a bit slower by maybe uh a meer 75% and it has a lot of waste by product where the hover car has no emissions but hey its still a better solution because its the dinosaur we know.
I ran into this years ago, and I would not be surprised if they are still doing it. But instead of upgrading their code to a better language they decided to upgrade Cobol by giving it all the features that the better more efficient languages have that Cobol did not have. And this was just the more simplistic constructs like While and For loops for example. So I would not be surprised if they have spent tons of money transforming Cobol into an object oriented language by now or are still working on doing so. Just so they have all that current technology within that dinotopic prehistoric language.
May 18, 2023 at 9:26 pm
Well, my workplace desktop support is useless on the SSC website issue. I am the only person in the office having problems with it but apparently the issue is SSC. And I sent an email the SSC webmster and haven't heard a thing back. So I have to access this website from my personal laptop, which is hard to read during work because it's twice as far from me as my work laptop. I'm dying here. I didn't realize how much of a lifeline this place was for me.
That being said, because of this I can't post screenshots properly, just text. And I urgently need help here: https://www.sqlservercentral.com/forums/topic/ssisdb-and-availability-groups-problem if someone has any thoughts on SSISDB and AGs and master key issues.
May 19, 2023 at 2:21 am
Yes, why would anyone want to use that complicated futuristic hover car when we can spend 10 times the amount of resources to build a better dinosaur. I mean, then we will still have the dinosaur and its simplicity. Yeah it is a bit slower by maybe uh a meer 75% and it has a lot of waste by product where the hover car has no emissions but hey its still a better solution because its the dinosaur we know.
You're insisting that it will be a dinosaur. I'll say "It depends". More times than not, someone has promised "futuristic hover cars" and, compared to the crap code that was written as a predecessor by someone that doesn't actually know SQL. it frequently is.
I try not to waste my time on "futuristic hover cars"... "transporters" work much better. 😉
Even with that, though, "It Depends". Sometimes a winged dinosaur that can fly at Mach 2 gets the trick done. Its only emissions are the occasional hover car that it ate along the way. 😀
--Jeff Moden
Change is inevitable... Change for the better is not.
May 19, 2023 at 2:55 am
Jeff Moden I did try to outline that in context. I was saying I have actually seen this before with a dynotopic prehistoric language (aka Cobol) where the powers to be were not biting the bullet and moving to a better language but they were instead trying to make their dynotopic prehistoric language more relevant. Which if you have ever coded in Cobol you will know that is next to impossible, which means expending at least 5 times the effort that it would take to change to a more viable language and still not guarantee you have a quality solution.
Further I was not saying someone was promising a futuristic hover car but that one (if not many) actually already existed and were being used. However, these archaic thinkers could not give up what they were using and embrace the futuristic change but felt it was worth wasting resources to great a dynotopic prehistoric solution which of course is going to fall significantly short of these futuristic hover cars.
Viewing 15 posts - 66,316 through 66,330 (of 66,547 total)
You must be logged in to reply to this topic. Login to reply