2

I am getting very weird interpreter behaviour:

>>> _
True
>>> type(True)
<class 'bool'>
>>> type(_)
<class 'bool'>

I tried this because _ came up as a suggestion in Bpython, but it seems to work in the normal interpreter too. I am using

Python 3.3.3 (v3.3.3:c3896275c0f6, Nov 16 2013, 23:39:35) 
[GCC 4.2.1 (Apple Inc. build 5666) (dot 3)] on darwin

Can anybody explain why _ is a substitute for True? Is it legacy, similarly to how ";" can be used to end commands, but is not necessary/encouraged?

EDIT: It seems to be random. This does not happen in a new terminal, but once I start working on something _ starts becoming true. What the hell is going on?

Azsgy
  • 3,139
  • 2
  • 29
  • 40

4 Answers4

7

_ will be the result of the last evaluated command - at interpreter start up there isn't any so you'll get a NameError... after that, you'll get the previous result... Try opening a new interpreter and doing 2 + 2... you'll see 4 returned, then type _... eg:

>>> _

Traceback (most recent call last):
  File "<pyshell#18>", line 1, in <module>
    _
NameError: name '_' is not defined
>>> 2 + 2
4
>>> _
4
Jon Clements
  • 138,671
  • 33
  • 247
  • 280
4
2 + 1
Out[19]: 3

_ + 3
Out[20]: 6

_ stores the last returned value. Try it out.

Stefan van den Akker
  • 6,661
  • 7
  • 48
  • 63
4

_ simply gives you the last result evaluated (in the REPL, not in an ordinary script). This can also mysteriously prevent objects from being deleted immediately.

Community
  • 1
  • 1
detly
  • 29,332
  • 18
  • 93
  • 152
1

_ in the interactive interperter is usually the last output you received.

>>> 1 + 1
2
>>> _
2

Note it only applies to outputs (printed data won't do).

yuvi
  • 18,155
  • 8
  • 56
  • 93