How Does String Slicing Work in Python?

A slice has the form text[start:stop:step]. The start is inclusive, the stop is exclusive, and the step controls the direction and the spacing.

text = "python"

print(text[0])        # p
print(text[0:3])      # pyt
print(text[2:])       # thon
print(text[:2])       # py
print(text[::2])      # pto
print(text[::-1])     # nohtyp
print(text[-2:])      # on
print(text[-3:-1])    # th

Reading the outputs: text[0:3] takes the first three characters, so pyt. Leaving the stop off means go to the end. Leaving the start off means start at zero. A step of 2 takes every second character, giving pto from the six-character string, and a step of -1 walks backwards, which is the standard way to reverse a string. The negative indices count from the end, so text[-2:] is the last two characters.

Two rules explain most slicing mistakes. First, the stop index is exclusive, unlike the inclusive ranges you might expect from an inclusive language. Second, an out-of-range slice does not raise an error — text[0:100] on a six-character string returns the whole string, and text[10:20] returns an empty string. Only indexing with a single out-of-range number raises IndexError, which is why a slice is safe to use on data of uncertain length.

The step being negative is also what makes the slice forms interchangeable with the common slicing idioms. text[::-1] is the reverse, text[:n] is the first n characters, text[-n:] is the last n, and text[1:-1] strips the first and last characters. Each of those is one operation with no intermediate copy beyond the result.

A slice selects a range of positions from a sequence. The start argument is the first position included, the stop argument is the first position excluded, and the step argument says how many positions to move between each one.

The exclusive stop is the convention that makes slicing composable. Because the stop is the first position you do not want, the length of a slice is simply stop minus start, and adjacent slices tile the sequence without overlapping or leaving a gap. text[0:3] and text[3:6] cover the whole string, which is what lets you process a string in fixed-size chunks.

The same convention is used everywhere else in Python — the range function has an exclusive stop, and slicing a list behaves the same way — so learning it once pays off in several places. It is also why len(text[:n]) is n and why the classic way to drop the last character is text[:-1] rather than text[:-0].

Three practical details. Omitting the start means zero and omitting the stop means the end of the sequence, so text[:] is a copy of the whole string. A negative index counts from the end, with -1 being the last character, and negative start and stop values are resolved against the length. And when the step is negative the slice walks backwards, which is what makes the reversal idiom work and also means text[::-1] is not the same as text[:0:-1] — the first includes the last character, the second does not.