SQL is still the master race with data manipulation/cleaning I've seen very few things that I see our data team using R or Python for that couldn't be done in SQL It's just not as fun
I mean...that's not really a fair comparison. By definition, the data is typically going to be cleaned and normalized before it can be stored in a relational database anyway. When I use R to clean data it's generally because it wasn't queried out of a database in the first place. Otherwise I generally just write SQL queries within my code to pull the data, and then leverage R functions to do the statistical analysis.
It can mean all sorts of things. Many times I'm working with XML or json, and parsing it to put it in tabular format. Other times it's log files that can be in any number of raw text formats. Sometimes it's SAR/Sysstat or Perfmon data captures from servers, while other times it's the output from some sort of home brew test harness a developer built, that outputs data in some amazingly stupid format. Other times it's just that the data isn't tidy (see this: https://cran.r-project.org/web/packages/tidyr/vignettes/tidy-data.html), and the built in packages in R (and probably Python pandas too), are just easier to use. Plus, we use like 6 different DBMS's at work, all with different capabilities and built-in functions for performing data cleaning and tidying tasks, and it's easier to just remember R syntax, than each individual database 's quirks.
jumping back on the C++ train using an app (Learn C++ by SoloLearn) even though I haven't done it in a while. Then gonna pick up Python via the same people I see they list calculus in that last tab. Just how useful is it in a programming context? Is it essentially just a first step in getting ready to read algorithm texts?
i've done everything in that sololearn thing without knowing any calculus calculus might be useful for something like cryptography if you're into that but i dont know for sure
I've used it for modelling moving stuff, but usually just by iterating rates onto values, or doing funky integrals with numerical methods. I know it can be used, I just don't know if there are more computer science-y and not necessarily computational context for it
Although it doesn't have much content on C++, the book "C Programming: A Modern Approach" is a really good read. It's pretty quick to get through it and helped me get a firmer grasp on what's going on in the compilers head after i send my shitty program to it.
Has anyone begun using the conda-forge channel to add packages? Trying to get an idea if it's worth doing that vs. using pip for packages that aren't in the standard Anaconda list.
I'm taking a computer systems class this summer. Today we learned about rounding errors in floating point numbers. The world just became a slightly scarier place.
ran into that in C earlier this year. it's so fucking finicky. also, checking in to say how much i love working with php.
also related to floating point rounding: apparently the plot of office space actually happened in the early days of computerized banking, more or less, which is why banks can't use floating point numbers when dealing with money stuff
You are so lucky you (presumably) are working with modern php, as in "object oriented is a thing!" php...
It's my fault, I drastically underestimated digits. For our code rounding error on 1/1000 of a pascal doesn't really matter. But when you are interested in fractions of a cent on trillions of dollars then you run up against double precision limit.
I try to now Once I was trying to pip install an Oracle python library. Wouldn't work because I was missing files it needed. Tried for a few hours to Google + install files it needed and still couldn't get it working . then tried conda install cx_oracle and it downloaded everything I needed and worked So team conda install
Yeah, I've started to use it. A lot of the geospatial/gis libraries were a bitch until i found conda forge
Currently working on a project that requires conversion from hex to 32-bit float, and I have to display the exponent bits, fraction bits, normalized/denormalized, +/- infinity, etc. How important is it that I learn this kind of thing? Visualizing hex->float-> binary so I can code this thing is pretty hard for me
numberphile has sister channel that does all computer stuff https://www.youtube.com/user/Computerphile/videos
Asked this on our board but figured this may be a decent place to put it too: I'm now in the market for a laptop. Looking for something affordable. I don't need anything really high performance, not gaming or anything. At most just using resource intensive applications like Tableau, but mostly internetting, listening to music, streaming video, etc. Thoughts on options or at least specs I should keep an eye out for?
looks like the lenovo L560 would be a good option for you. they have options for i3 and i5, but i'm not totally sure how processor dependant tableau is. i'd usually recommend i5 or higher for productivity. http://shop.lenovo.com/us/en/laptops/thinkpad/l-series/l560/ i3 is $719, and the i5 is $804. might be better off looking at newegg or microcenter though.
learned about buffer overflow today i'm basically a hacker now this is in the instructor's lecture notes: By the way, you are being taught about this so that you can defend yourself from attacks. If I ever hear that one of my students has used these techniques for evil, I will personally come and testify against that student at their trial, even if I have to spend my own money to get there.
currently learning how to do parent/child process stuff i think i understand the way that tmb was attacked that one time it was a fork bomb
And is completely pointless, too I think they only teach it to separate the regular nerds from the hard core aspies and savants.
The guiding logic of the program I'm in is more or less that we have to learn shit like it's the 1980s because local industry said that today's graduates have weak fundamentals. The guy teaching the class was like "there are optimizing compilers that do all this stuff for you now. you still have to do this project." I guess it's useful for learning about the stack.
Any of y'all used python to query a mysql table and insert it into another mysql or maybe just a csv files? Seems pretty straight forward
I Do all the time with teradata, Oracle, and SQL server The documentation is pretty simple You can convert it to a pandas data frame easy too if needed
What kind of program are you in again? Sure, everyone now-a-days has weak fundamentals because you learn that it's not important, and in fact a waste of time, to worry about trying to write your own implementation of a red-black tree or a trie.
And the reason today's industry says "graduates have weak fundamentals" is so they can convince congress to bring over cheaper labor from asia on work visas.
Computer science. I'm taking the lower division classes at a community college because $, but the curriculum is dictated by Portland State because like 95% of the people at my school transfer there. It's funny you mention that last part, because in order to transfer you have to pass a proficiency exam in which you draw an index card from a bowl that will have something like "red-black tree" on it and you have to write it with no internet/book, in vim or emacs with their rc files (ie, no arrow keys and definitely no mouse), and people looking over your shoulder as you do it
Yeah that seems a bit much, way too much of an old school focus for nothing more than the sake of being old school. I didn't graduate with a CS degree, but I took a two online courses for Algorithms and Data Structures that were designed by a couple of professors at Princeton (one of which designed the Red-Black tree). They had a lot of assignments where they required you to apply knowledge of these structures to solve a bigger problem like you would in real life (this was very challenging and very rewarding). It's extremely rare that you'll need to design something so low level and fundamental, but it is extremely common that you know how to use such things.
and I hate any time you're asked to do something using tools/resources other than you would actually use if you were "on the job" or whatever. Like doing code review with printed sheets of shit instead of having it on a screen with line numbers, folding, etc. FOH.