Starting from version 9.0, the default distributed version of Sage is using Python 3. See Python3-Switch for more information.
Main caveat from writing Python 3 Sagemath code when you are used to Python 2
We list the most visible changes that any user would certainly run into the first time she uses a Python 3 based version when she used to run a Python 2 based one. This page is not intended to be a complete manual of the differences but a short page of what you need to know to get started. The What's new page of the Python official website provides more technical details.
In Python 2, print is a keyword (as for, if, etc)
In Python 3, print becomes a function
Writing a print statement without the parenthesis results in a SyntaxError
One cool thing about this print function is that it takes various arguments (you can write in files or use a different separator).
2. range and xrange
In Python range is a function that returns a list.
In Python 3, range is an object that somehow behave as a list (ie elements can still be acessed with square bracket, it has a length, etc) but it is not a list.
The main advantage of this range object in Python 3 is when you perform a simple iteration for i in range(1000). In Python 2 this would have created a list of 1000 elements and then iterated through it. In Python 3, there is no list creation and hence no memory allocation for these 1000 elements.
The iterator xrange is no longer valid in Python 3 simply use range instead.
3. strings, unicode and bytes
In Python 2 a chain of characters between simple, double or triple quotes creates an ascii string. To create a unicode string one has to use the prefix u (which remains valid in Python 3).
In Python 3, a chain of characters between quote will gives rise to a unicode string. To create specifically an ascii string one has to use the prefix b (which is already valid in Python 2).