We recently picked up python, mostly by accident. We had some really shorty formatted XML. We’re talking no-schema because nodes of the same type contain different elements. No consistency. Some nodes used differently from each other depending on the purpose and three different model-trees dropped into one giant mess, not at random but at datapoint entry but without the time stamp.<p>C# just couldn’t handle it, none of the libraries we tried could and when we wrote our own it was just hopelessly inefficient. Enter python and everything was as simple as telling python which element value/text/whatever they decided to store the data in and you’d get it. 200 lines of code later and a 5 second runtime we had our data in neatly store objects in a couple of dictionaries.<p>At that time we could have dumped it onto our SQL cluster and let C# take over, but the programmer in charge decided to just go at it with python. Mostly for the heck of it. A few weeks later he had replaced a major SSIS service and multiple C# services with some relatively simple python that cut the total runtime from 30 minutes to 5.<p>I’d never written a line of python before this, but reading the code, I understood everything to the point where I’d feel comfortable maintaining/expanding it, with very minuscule help from the documentation (and google).<p>I don’t personally have that much love for dynamic types, and I’m not sure how we would use it, if we needed to run it outside our environment, but the experience perfectly highlights how python won and repeatedly continues to win our hearts.