如何在python中将多个函数的输出重定向到字符串或标准输出？ （stdout 到变量的反向）

Question

Often I have to print all sorts of diagnostics and I am tired of wrapping everything into a print. Is there something that allows me to just redirect output ie return values of multiple function calls to a string or stdout?

import numpy as np
import pandas as pd

df = pd.DataFrame({'a':range(10)})

# tedious
print(df.describe())
print(df.head())

# better
with something():
    df.describe()
    df.head()

I am aware of Redirect stdout to a file in Python? that is not what I want to do.

Answer 1

I assume that you want to print the return values of those functions (functions don't "have output" unless they themselves use print or write to a file, in which case your problem would already be solved). You could do this:

def print_many(*values):
    for value in values:
        print value

print_many(
    df.describe(),
    df.head(),
)

Note that this will execute all the functions before anything is printed.

Answer 2

Short answer: no.

Long answer: you can maybe do something like this:

def print_all(*expressions):
    for expr in expressions:
        print(expr)

print_all(
    df.describe(),
    df.head(),
)

which comes with a couple caveats:

1. You couldn't use conventional assignment. When assignment would be required, you'd have to use the assignment operator := , which is only available in python 3.8, so this would be structured differently than a normal function.
2. Everything would be evaluated first, before being printed. You could work up some dynamic thing probably by using eval() , but that has other risks.

---

Alternatively, you could try modifying the df object beforehand so that all relevant methods print their return values - essentially, applying a decorator to every method you want:

def print_wrapper(func):
    def wrap(*args, **kwargs):
        retval = func(*args, **kwargs)
        print(retval)
        return retval
    return wrap

def wrap_funcs(obj, funcnames):
    for funcname in funcnames:
        func = getattr(obj, funcname)
        setattr(obj, funcname, print_wrapper(func))

df = pd.DataFrame({'a':range(10)})
wrap_funcs(df, ['describe', 'head'])

df.describe()
df.head()

This also has caveats, as you'd need to ensure that the names you give it actually are functions or things start going wrong fast, but it should be a possible starting point.