1, Python performance, I think there is no need to worry about the needs of most Internet companies. Someone in Zhihu is using Python to do high-frequency trading VNPY/VNPY GitHub. If you really want to deduct performance, write the core module in pure C and package it for Python to call.

2. For the robustness of engineering development, Python is not as good as Java. After all, Java is a language that is good at making small projects into big projects. From this point of view, if you think your project can set the general direction in one breath, and it will remain unchanged for several years, it is recommended to use Java. After all, the open source framework uses more people and will be relatively reliable; If you are a start-up company, or haven't set foot on formal business for a long time, the demand will often be a few days small and a few weeks long. At this time, the thinking of agile development and lean entrepreneurship is particularly important, and the advantages of Python will be greatly exerted.

Having said that, back to the problem itself, the main reason is that it is too difficult to find Pythoner in China. As far as my personal experience in recruiting fresh graduates is concerned, the written test and interview are mainly about C++ and basic algorithm knowledge, and then we will teach Python ourselves after recruiting. Unfortunately, most recruiters say that they have never been exposed to Python and are not very interested. The proportion of Pythoner recruited by social organizations is even smaller, and Javaer basically accounts for half of the country. Therefore, considering the mobility of talents and the cost of employing people (code farmers love job hopping), many companies have no choice but to switch from Python to Java.