BUG: "A value is trying to be set on a copy of a slice from a DataFrame" even if I set something on a copy dataframe

[GF-Huang]

Pandas version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• I have confirmed this bug exists on the main branch of pandas.

Reproducible Example

I can't, it's complex.

Issue Description

I already copy, but it still say A value is trying to be set on a copy of a slice from a DataFrame.

Expected Behavior

No warning if I copy the df then set something .

Installed Versions

INSTALLED VERSIONS

---

commit : 66e3805
python : 3.9.1.final.0
python-bits : 64
OS : Windows
OS-release : 10
Version : 10.0.19041
machine : AMD64
processor : Intel64 Family 6 Model 94 Stepping 3, GenuineIntel
byteorder : little
LC_ALL : None
LANG : None
LOCALE : Chinese (Simplified)_China.936

pandas : 1.3.5
numpy : 1.19.5
pytz : 2021.1
dateutil : 2.8.2
pip : 21.1.3
setuptools : 57.4.0
Cython : None
pytest : None
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : 4.6.3
html5lib : None
pymysql : 1.0.2
psycopg2 : None
jinja2 : 3.0.1
IPython : 7.25.0
pandas_datareader: None
bs4 : None
bottleneck : 1.3.2
fsspec : None
fastparquet : None
gcsfs : None
matplotlib : 3.4.2
numexpr : 2.7.3
odfpy : None
openpyxl : 3.0.9
pandas_gbq : None
pyarrow : None
pyxlsb : None
s3fs : None
scipy : 1.7.0
sqlalchemy : 1.4.22
tables : 3.6.1
tabulate : None
xarray : None
xlrd : 2.0.1
xlwt : None
numba : None

[phofl]

Please provide something reproducible, otherwise we are not able to investigate. Your screenshots are not sufficient

[phofl]

Please provide an example that is copy and pastable.

[GF-Huang]

Data file: data.zip

import pandas as pd

df = pd.read_csv('data.zip', compression='zip')
df.sort_values(['TimeBarStart', 'Strike', 'CallPut'], inplace=True)

open_price = (df.iloc[0].UnderOpenBidPrice + df.iloc[0].UnderOpenAskPrice) / 2

df = df[df.Strike == df.loc[(df.Strike - open_price).abs().idxmin()].Strike]
df.set_index('TimeBarStart', inplace=True)
df.index.rename('Time', inplace=True)
call, put = df[df.CallPut == 'C'][:'15:59'], df[df.CallPut == 'P'][:'15:59']
straddle = (call[['OpenBidPrice', 'HighBidPrice', 'LowBidPrice', 'CloseBidPrice', 'Volume']] + 
            put[['OpenBidPrice', 'HighBidPrice', 'LowBidPrice', 'CloseBidPrice', 'Volume']])
straddle['UnderOpen'] = ((call.UnderOpenBidPrice + call.UnderOpenAskPrice) / 2 + \
                          (put.UnderOpenBidPrice + put.UnderOpenAskPrice) / 2) / 2
straddle['UnderClose'] = ((call.UnderCloseBidPrice + call.UnderCloseAskPrice) / 2 + \
                          (put.UnderCloseBidPrice + put.UnderCloseAskPrice) / 2) / 2
call_spread = call.CloseAskPrice[0] - call.CloseBidPrice[0]
put_spread = put.CloseAskPrice[0] - put.CloseBidPrice[0]
straddle = straddle.copy()
straddle.CloseBidPrice[0] = (call.CloseBidPrice[0] + call_spread * 0.5) + (put.CloseBidPrice[0] + put_spread * 0.5)
straddle.OpenBidPrice[1] = straddle.CloseBidPrice[0]

[phofl]

Is all this necessary? Can you remove Parts of this and it shows still the error? Please remove everything that is not necessary to reproduce the bug. Copy and pasteable means that it should not depend on files if possible. Viewing your screenshot, this should be possible.

[GF-Huang]

The only differences is the Volume column.

import pandas as pd

df = pd.read_csv('data.zip', compression='zip')
df.set_index('TimeBarStart', inplace=True)
call, put = df[df.CallPut == 'C'], df[df.CallPut == 'P']
straddle = (call[['OpenBidPrice', 'CloseBidPrice', 'Volume']] + 
            put[['OpenBidPrice', 'CloseBidPrice', 'Volume']])
straddle = straddle.copy()
straddle.OpenBidPrice[1] = straddle.CloseBidPrice[0]

[phofl]

Thx, this is better. Can you create the dataframe in code instead of reading it from a file? Not more rows than strictly necessary please

[GF-Huang]

I can't, you can try download my uploaded data.zip, it very large (4MB).

[phofl]

The probability that someone takes a closer look increases, if it is easier to replicate. Large dataframes are terrible to debug.

[GF-Huang]

I'm sorry, I can't do any more.

[jorisvandenbossche]

This is quite easily reproducible with a small example like this:

In [11]: df = pd.DataFrame({"a": [1, 2, 3], "b": [3, 4, 5], "c": [0.1, 0.2, 0.3]})

In [12]: df.a[0] = df.b[0]
SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  df.a[0] = df.b[0]

I suppose we could consider this as a false positive (modifying a column like this should I think in theory always work?), but you are still doing "chained assignment" (assigning something to a certain column and certain row label, but selecting column and row in separate, chained steps). Generally, the recommendation is to use .loc/.iloc instead to do the assignment in one step. For my small example, this is:

In [13]: df.loc[0, "a"] = df.b[0]

which doesn't raise the warning.

In your case, you have a string row index, so if you want to combine a positional row label with a named column label, you might need to do something like straddle.loc[straddle.index[1], "OpenBidPrice"] = ....