[tannk11001]

As an engineer that identifies as being self-taught, even while having done some formal CS education after I'd already started my career, this question strikes me as unclear.

As far as I've seen in my own code, and those of my coworkers, we use all kinds of data structures when they're appropriate to a requirement and its constraints.

Are you suggesting that some people stubbornly don't use data structures that they're familiar with, even when they're important to speed/performance? I can't say I really see that happen very often. It's usually either that the engineer didn't happen to know the optimal solution, or (as was mentioned elsewhere) the optimal solution wasn't necessarily worth pursuing.

[nvbhargava]

Yes, I was wondering if people stubbornly don't use data structures that they're familiar with. To be more clear, I meant implementing a whole new data structure (let's say a trie) that a library/language might not already provide.

This, of course, is taking into consideration that the problem is non-trivial, and it improves the speed/performance.