re: "Learn to Code"

quote:

---

SQL is also much easier than learning a regular language like Java or C++. Much easier

---

Probably gonna have to pair SQL with Python or something to do anything useful.

quote:

---

SQL can be extremely hard, the people getting employed with just sql skills, are doing very complicated stuff.

---

Yeah it gets convoluted real quick if you go past queries and start using procedural extensions. I honestly have no idea how deep that hole goes. I've only looked far enough to decide F that and always use something else to massage data.

Not that I do anything really worth a shite, mostly hobby stuff. I guess there are use cases where all that is a good fit because someone invented it and people use it, but using a programming language instead seems way easier to follow and maintain.

quote:

---

There is a meme among programmers, people who get paid to code, that their most important skill is googling for code samples. Impostor syndrome is rampant. Don't get discouraged.

---

Felt good that I did my first “solo” treehouse “project” yesterday.

They told me to create a number guessing game with a few parameters. No code given.

I was able to code it up in less than an hour. On one hand it’s simple, and basic. On the other hand, for someone who swore off coding in 2016 with my failed attempted to learn Java it felt amazing. I’ve only been at this for a week.

As an IT person I feel super comfortable with the whole google stuff and then implement it. Example. The course never taught anything about generating random numbers or storing that value, nor anything about incrementing a value(attempts at guessing correct random number).

But with a little bit of Google fu I figured it out quickly.

Really pumped to continue this journey. Because adding python to my already pretty extensive networking and Cisco skills should make me more marketable. It feels like there is an under 100k ceiling on network jobs without knowing python these days. At least around here.

quote:

---

There is a meme among programmers, people who get paid to code, that their most important skill is googling for code samples. Impostor syndrome is rampant. Don't get discouraged.

---

Just another skill and largely based in truth honestly. If you're a "good googler", you can provide quick solutions for many issues that seem like they could take months to get to the root cause.

The mistake is when I see people think coding equals programming. What you need to do is learn programming and forget the syntax when you first start. Learn the fundementals: Conditional logic, loops, and variables. Then move on to object oriented concepts. Once you understand the basic logic behind all those concepts, you can program in any RAD programming language today: C# is probably the most popular right now.

I mean really understand these things from a logic stand point; do not worry about coding yet. Once you know all those things, then select a programming langue and start coding. Find a small project you want, something simple, maybe adding up a few numbers, really simple. Then select a language and google is your friend for syntax.

I have been in coding and system administration for 30 years. I have three degrees in technology and have taught it for years as well as been in practical IT. On a daily basis, I rinse and repeat those basic concepts. You do not need a formal degree, you do need to understand computer logic and have the ability to break down problems into their components and then build solutions to address those problems. That is really what programmers do all day long, coding is just the last step.

FYI, do not learn SQL before you learn programming, it will mess you up because SQL is very specific logic for relational databases. There are other types of data structures. Once you understand computer programming, then start learning data structures and databases. Most programmers make money pulling data out of databases for reports and other data mining activities.

quote:

---

For a career, if you're over 40 don't waste your time. Ageism in IT is rampant.

---

This is true if you havent updated your skills in 10-20 years. If you keep up to date with technology and continue to learn, this is far from the truth.

quote:

---

For a career, if you're over 40 don't waste your time

---

I disagree with this do not let your age be a barrier. This is simply not true.

quote:

---

honestly with how capable the Python data ecosystem is now a days. Learn basic SQL, use SQL as a data store and get the data into Python as quickly as possible.

---

I still argue this point to no end. Python is a scripting language, not an ETL tool. If you need to massage the data why not use a tool designed to do that?

quote:

---

I still argue this point to no end. Python is a scripting language, not an ETL tool. If you need to massage the data why not use a tool designed to do that?

---

I can only speak from a data science perspective, and not from a more hardcore dev perspective. However, I'd argue pandas is a tool specifically designed for ETL and is damn good at it.

As to why I use pandas for ETL vs SQL. If I'm web scraping or hitting an API to populate a DB, I'm using Python. Training machine learning models, in Python again. Building an API to serve predictions, still Python.

You say scripting language as a negative. However, for data science all we do is basically write scripts. Want to turn those scripts into a more reliable workflow, use a data pipeline orchestration tool like Airflow, with you guessed it, Python.

From a data perspective, nothing comes close to Python in the breadth of tools available IMO.

I have nothing against python, there are cases where scripting makes sense. Machine learning, yeah that's a python thing.

Hitting an API to populate a db? That's an ETL tool thing. Pentaho, informatica, ssis. All have built in capability to point to an API and grab the data in a few clicks.

It's a case by case thing, there are reasons to use either and we do use either. Just use the tool that makes sense to use, not just because it can do it, reason being maintenance and performance.

It's the same argument with Excel, yes Mr business user I'm very aware that you *can* do that in Excel. Is it this best tool for the job though?

ETA: our data scientists use python and Pentaho. Our warehousing team uses Pentaho, which supports Java and Python