Part 1 uh can't you just split it into an array and do comparison?

I hate interview questions like this, seems like it'd be very loosely correlated with whether someone's actually a good programmer or not.

I wouldn't recall the exact equation off my head, but this is how I'd do it....



With goofy on this one. I hate these kinds of questions.

dont you need to turn it into a string before you can split it into an array?

oh I guess you can mod by 10 for each index.

so
List=digits
while(n>0)
digits.add(n%10)
n=n/10

and now that I start thinking about this, I actually had this question for an interview. I got it right lol. I used to be better at this stuff.

Yup. I'm more and more convinced that a screening question like this should have a pretty obvious solution (or better, solutions) and the candidate just has to show that they can take that abstract solution and make it work in code.

Is there a reason other than didn't think of it/doesn't matter why your is_palindrome function isn't just the 1 line: "return n == reverse_int(n)"? (Honest question, just making sure I understand the ins and outs of python)

Does this count? I'm not comparing 2 arrays, and technically I'm not converting my original integer to a string, just my reverse array.

Yea those are good, I worded the second half wrong.

The second half is doing it without comparing anything to the reverse of it.

(I think the solution is kinda cheesy)

Solution being asked for:

I've been learning SQL for the last 24 hours and I have a new appreciation and empathy for people who have screwed up production databases before.

I can't even change things on my local without nervously checking select and being careful that I know exactly what I'm doing.

Brag, I got the job! obviously delighted, means that I won't have to bother with job search anymore <3

The question asked to not convert to string.

There is nothing wrong with your suggestion, and I could argue that it is indeed better and more pythonic.

I have a habit of writing things with more verbosity than needed on a first write. I'm also more lispy, so I tend to think in let bindings.

well if you can figure out the length then you can just use math to reconstruct the reverse by %10 and then compare the 2.

ok so if n is the number to check.



and Im sure there is way to only loop once.

"begin;

do your SQL here

-- commit;
-- rollback;

-- uncomment commit or rollback that you want to use you want to use.
"

Manual transactions are your BFFs.

Yeah that's pretty cool. I knew there had to be something easy to figure out the # of digits. But I felt like googling "how to determine the number of digits" might be cheating.

Congrats!!!

And Larry, here is another trick that is super awesome. Not many people know it, but if you have a result set you like and you want to check if your rewrite is accurate, or just see if you are making a mistake on either the initial query or new query:


If you end up with the empty set, you have matching results. If you have results, then you know there is a mistake somewhere. By default union and except remove duplicates. Using "all" returns duplicates.

If you are using PostgreSQL, you should be able to find a procedure for this online. That way you can just do something like "select matching(view_a, view_b);"

awesome, sounds like a sweet opportunity

had another candidate come in, same question, did it in about 20 minutes without any outside references

Apparently my friend just got her name on this patent: http://patft.uspto.gov/netacgi/nph-P...F9%2C635%2C195

Which as far as I can tell is just a generic interface to view video content. Lol tech patents always and forever.

Tech patent law is some of the dumbest stuff I've seen. It's an area I considered getting into after I graduated from college, and I am so glad I didn't. I have a friend who worked for a company that primary did statistical analysis to be used as expert testimony in court cases. A lot of their work was on patent cases. So much fail all around.

I did the same thing as my first real programming job (SAS, stata, perl) but most of our work was on tobacco and auto lawsuits. Your friend didn't work for a company in Novato, CA did he?

Nope, Dallas, TX. One of the premiere venues for IP law cases is in a podunk district of Texas. This is because the district is very favorable towards (one side) of IP law disputes.

https://www.eff.org/deeplinks/2014/0...as-its-not-bbq

BTW he now does statistics for a major medical device manufacturer and holy **** do the stories he tells me make me suspicious of medical research.

I can guarantee you from working for the pro-tobacco side (yeah, big job satisfaction) that the researchers' studies on tobacco harm we had to counter were often completely full of holes and made up bull****.

This one guy Harrision from Harvard had a reputation as the worst of the bunch. He produced some study which showed a certain curve for some instance of disease vs. smoking prevalence. He provided no supporting data, except to point to the NIH data we all used. My boss gave it to me and said "see if you can replicate this curve".

The curve I got, as he described in his methodology, looked nothing like the pic in his paper. So I started trying different mistakes he could have made - getting closer but still not matching. After two weeks I finally cracked the code. By introducing 3 or 4 different mistakes in the right combination - I perfectly matched his chart.

To this day it remains my proudest programming accomplishment. I burst into my boss' office with the good news, and explained the whole thing to him. He said "that's great but unfortunately it's way too complicated, a jury will never understand it". If Harrison had just made one simple mistake that we could demonstrate, instead of several - it would have been much worse for him. Ugh.

Yeah I've heard that before from my friend. The truth is not your ally, unless it's a story you can easily sell to a jury or a (non mathematical) judge. It's actually easier to sell a false pat story than a true complicated one. What can you do. I have literally heard your exact story, down to trying to replicate what "mistakes" a person did to turn the raw data into their conclusions.

I mean, it's not a coincidence that the tobacco industry employs a statistician who makes so many mistakes that are helpful to their cause.