It is well-known among Python developers (or at least, it should be), that is a bad idea to import everything from a module, for example by doing from <module> import *.
The idea on this post, is to highlight and gather the reasons why this is a bad practice, in order to collectively identify many undesired effects. However, without losing pragmatism, in case there are some odd reasons why this might be acceptable, I will also mention them, if any. Let’s see where we get.
- You do not know what you get
- An arbitrary Python script may contain any code, and most of it will be executed when performing the import * part (you cannot rely on how __name__ is handled). The interface is totally unclear: you do not know what computations performs, what objects will import, etc. In general is more efficient to import as few definitions as possible.
- Identifiers appear magically
- In any decently readable Python script, the programmer must be able to locate every definition, which means to identify where does every identifier come from. For example, a variable named x can either be a parameter of the function in the current scope, a variable already defined (assigned), or a name already imported (from mod import x), etc. By performing the incorrect import, this variable might appear out of the blue, meaning that I will have an x that will not be neither a parameter, nor a definition nor a declared import. This means, I cannot track the origin or x. The situation gets worse if there are not one, but many import * statements. Debugging becomes a nightmare.
- Namespaces are one honking great idea — let’s do more of those!
- Straight from the Python zen . By importing everything from a module, the benefits of the namespaces are somehow lost. Instead, everything (or a lot of things), might get to be called the same, messing with the current scope. Moreover, new import definitions might override previous ones.
- Explicit is better than implicit
- Again, every identifier that we want imported should be done explicitly (the * is not very declarative).
Now, so far these might be some of the main reasons about why importing everything from a Python module is usually not a good idea. However, in case the code at stake is just a simple testing script, or an in-line sentence on ipython, there could be nothing wrong about it.
In addition, although I am not a big fan of import statements inside functions (sometimes they are necessary, though), importing everything from a package within a function is not a big problem, because the scope is already narrowed.
Just to be clear, this is by no means an absolute statement, but an idea presented in order to write better code. One of the things I like the most about Python is that encourages good practices. Therefore, if I read an import statement like this, unless there are some very good reasons to do so, I will think that line as a code-smell .