Clip Could take Series/DataFrames

[rhaskett]

I was a little surprised when clip would not take other pandas objects as parameters. I hacked a little something below, but for readability I thought I might suggest that this feature be added.

df1 = DataFrame(ones((2, 2)))
data = ones((2, 2))
data[0, 0] = 0
data[1, 1] = 2
df2 = DataFrame(data)
print df2.clip(1.5, 0.5)

df3 = df2.copy()
df3[df2 > df1 * 1.5] = (df1 * 1.5)[df2 > df1 * 1.5]
df3[df2 < df1 * .5] = (df1 * .5)[df2 < df1 * .5]
print df3

[brandon-rhodes]

I was also surprised, this morning, that clip() does not accept Series as bounds for pair-wise comparison. Here is an example of my current work-around, compared with how I thought that clip() would operate (in case it makes the issue clearer for anyone, since the issue above does not actually show how clip() ought to work):

import pandas as pd
from StringIO import StringIO

df = pd.DataFrame.from_csv(StringIO("""\
,over,but_not_over,rate
0,692.00,2179.00,0.10
1,2179.00,6733.00,0.15
2,6733.00,12892.00,0.25
3,12892.00,19279.00,0.28
4,19279.00,33888.00,0.33
5,33888.00,38192.00,0.35
6,38192.00,inf,0.396
"""))

wages = 29279.00

# Today, I have to write:

wages_in_bracket = (df['but_not_over'].clip_upper(wages) - df['over']).clip_lower(0.00)
print wages_in_bracket

# But I had expected to be able to:

wages_in_bracket = df['but_not_over'].clip(df['over'], wages) - df['over']
#
# Traceback (most recent call last):
#  ...
# ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

[kokes]

Isn't this a duplicate of #6966? I've just tested the second example and it does work.

[jreback]

yep

[brandon-rhodes]

I was commenting on Pandas issues back in 2013 — five years ago? Wow, I've used Pandas for a long time!

I'm happy to hear that clipping against series is now possible!