英文:
'float' object has no attribute 'rint'
问题
DataFrame的所有单元格都是浮点型，但仍然无法四舍五入。
DataFrame
数据类型
我正在执行以下操作：
df_lmsd['LMSD']=np.around(df_lmsd['LMSD'],2)df_lmsd['LMSD(#)']=np.around(df_lmsd['LMSD(#)'],2)df_lmsd
输出结果：
Python版本：Python 3.10.9
IDE：Jupyter Notebook
英文:
All the cells of DataFrame are of float type but still it is not able to round off.
DataFrame
Dtype
I am doing this:
df_lmsd['LMSD']=np.around(df_lmsd['LMSD'],2)df_lmsd['LMSD(#)']=np.around(df_lmsd['LMSD(#)'],2)df_lmsd
Output:-
Python Version:- Python 3.10.9
Ide:- Jupyter Notebook
答案1
得分: 3
存储类型为object的LMSD列，而不是float。为了演示这一点，我已经重现了错误。
输入数据（将LMSD列存储为float类型）
import pandas as pdimport numpy as npdata = {"created_date": ["2023-03-01", "2023-03-03", "2023-03-07", "2023-03-08", "2023-03-09"],"lmsd": [99.701135, 100.0, 4.142565, 14.182277, 24.433356],"lmsd_hash": [99.853902, 100.0, 4.579792, 14.58302, 23.937297]}df = pd.DataFrame(data)# 将日期转换为日期时间对象df["created_date"] = pd.to_datetime(df["created_date"])print(df)
created_date lmsd lmsd_hash0 2023-03-01 99.701135 99.8539021 2023-03-03 100.000000 100.0000002 2023-03-07 4.142565 4.5797923 2023-03-08 14.182277 14.5830204 2023-03-09 24.433356 23.937297
df.info()的输出:
<class 'pandas.core.frame.DataFrame'>RangeIndex: 5 entries, 0 to 4Data columns (total 3 columns):# Column Non-Null Count Dtype--- ------ -------------- -----0 created_date 5 non-null datetime64[ns]1 lmsd 5 non-null float642 lmsd_hash 5 non-null float64dtypes: datetime64[ns](1), float64(2)memory usage: 248.0 bytes
将LMSD列存储为object类型
# 将LMSD列表示为对象df[["lmsd", "lmsd_hash"]] = df[["lmsd", "lmsd_hash"]].astype(object)print(df)
created_date lmsd lmsd_hash0 2023-03-01 99.701135 99.8539021 2023-03-03 100.0 100.02 2023-03-07 4.142565 4.5797923 2023-03-08 14.182277 14.583024 2023-03-09 24.433356 23.937297
（请注意，这与您的DataFrame匹配）
df.info()的输出:
<class 'pandas.core.frame.DataFrame'>RangeIndex: 5 entries, 0 to 4Data columns (total 3 columns):# Column Non-Null Count Dtype--- ------ -------------- -----0 created_date 5 non-null datetime64[ns]1 lmsd 5 non-null object2 lmsd_hash 5 non-null objectdtypes: datetime64[ns](1), object(2)memory usage: 248.0+ bytes
尝试在object列中四舍五入
# 尝试在LMSD列中（表示为对象）四舍五入值df['lmsd'] = np.around(df['lmsd'], 2)df['lmsd_hash'] = np.around(df['lmsd_hash'], 2)
会产生以下错误，与您收到的错误相匹配:
TypeError: loop of ufunc does not support argument 0 of type float which has no callable rint method
在四舍五入之前将列表示为float
为了解决这个问题，确保在四舍五入之前将LMSD列存储为float，使用astype()方法:
# 将LMSD列转换为float值。df[["lmsd", "lmsd_hash"]] = df[["lmsd", "lmsd_hash"]].astype(float)# 将值四舍五入到小数点后两位df['lmsd'] = np.around(df['lmsd'], 2)df['lmsd_hash'] = np.around(df['lmsd_hash'], 2)print(df)
输出:
created_date lmsd lmsd_hash0 2023-03-01 99.70 99.851 2023-03-03 100.00 100.002 2023-03-07 4.14 4.583 2023-03-08 14.18 14.584 2023-03-09 24.43 23.94
英文:
It appears that you have the LMSD columns stored as the type object rather than float. To demonstrate this, I have reproduced the error.
Input Data (with the LMSD columns stored as type float)
import pandas as pdimport numpy as npdata = {"created_date": ["2023-03-01", "2023-03-03", "2023-03-07", "2023-03-08", "2023-03-09"],"lmsd": [99.701135, 100.0, 4.142565, 14.182277, 24.433356],"lmsd_hash": [99.853902, 100.0, 4.579792, 14.58302, 23.937297]}df = pd.DataFrame(data)# Convert the dates to datetime objectsdf["created_date"] = pd.to_datetime(df["created_date"])print(df)
created_date lmsd lmsd_hash0 2023-03-01 99.701135 99.8539021 2023-03-03 100.000000 100.0000002 2023-03-07 4.142565 4.5797923 2023-03-08 14.182277 14.5830204 2023-03-09 24.433356 23.937297
Output of df.info():
<class 'pandas.core.frame.DataFrame'>RangeIndex: 5 entries, 0 to 4Data columns (total 3 columns):# Column Non-Null Count Dtype--- ------ -------------- -----0 created_date 5 non-null datetime64[ns]1 lmsd 5 non-null float642 lmsd_hash 5 non-null float64dtypes: datetime64[ns](1), float64(2)memory usage: 248.0 bytes
Storing the LMSD columns as type object
# Represent the LMSD columns as objectsdf[["lmsd", "lmsd_hash"]] = df[["lmsd", "lmsd_hash"]].astype(object)print(df)
created_date lmsd lmsd_hash0 2023-03-01 99.701135 99.8539021 2023-03-03 100.0 100.02 2023-03-07 4.142565 4.5797923 2023-03-08 14.182277 14.583024 2023-03-09 24.433356 23.937297
(Notice how this matches your DataFrame)
Output of df.info():
<class 'pandas.core.frame.DataFrame'>RangeIndex: 5 entries, 0 to 4Data columns (total 3 columns):# Column Non-Null Count Dtype--- ------ -------------- -----0 created_date 5 non-null datetime64[ns]1 lmsd 5 non-null object2 lmsd_hash 5 non-null objectdtypes: datetime64[ns](1), object(2)memory usage: 248.0+ bytes
Attempting to round object columns:
# Try to round the values in the LMSD columns (represented as objects)df['lmsd'] = np.around(df['lmsd'], 2)df['lmsd_hash'] = np.around(df['lmsd_hash'], 2)
The following error is produced, matching the one you were given:
<pre>
TypeError: loop of ufunc does not support argument 0 of type float which has no callable rint method
</pre>
Representing the columns as float before rounding
To fix this, ensure that the LMSD columns are stored as floats before rounding, using the .astype() method:
# Convert the LMSD columns to float values.df[["lmsd", "lmsd_hash"]] = df[["lmsd", "lmsd_hash"]].astype(float)# Round the values to two decimal placesdf['lmsd'] = np.around(df['lmsd'], 2)df['lmsd_hash'] = np.around(df['lmsd_hash'], 2)print(df)
Output:
created_date lmsd lmsd_hash0 2023-03-01 99.70 99.851 2023-03-03 100.00 100.002 2023-03-07 4.14 4.583 2023-03-08 14.18 14.584 2023-03-09 24.43 23.94
答案2
得分: 0
你或许可以这样做：
import numpy as npimport pandas as pddf_lmsd = pd.DataFrame({'LMSD': [1.23456789], 'LMSD(#)': [2.34567890]})df_lmsd['LMSD'] = np.around(np.array(df_lmsd['LMSD']), 2)df_lmsd['LMSD(#)'] = np.around(np.array(df_lmsd['LMSD(#)']), 2)print(df_lmsd)
这将会得到：
LMSD LMSD(#)0 1.23 2.35
英文:
You should maybe do this:
import numpy as npimport pandas as pddf_lmsd = pd.DataFrame({'LMSD': [1.23456789], 'LMSD(#)': [2.34567890]})df_lmsd['LMSD'] = np.around(np.array(df_lmsd['LMSD']), 2)df_lmsd['LMSD(#)'] = np.around(np.array(df_lmsd['LMSD(#)']), 2)print(df_lmsd)
which gives
LMSD LMSD(#)0 1.23 2.35</details>
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论