主页 / coding / 问题 / 79058570
Accepted
rakesh
Asked: 2024-10-06 15:31:54 +0800 CST

Pandas 获取底部第一个负数

  • 772

我有以下 df。其中我想看到最后一个结果数字之后的第一个负数。＆在“最后一个负数”列中“这应该出现，或者更确切地说是任何突出显示该行的内容。“这应该出现在最后一个结果数字之后的第一个负数上（结果数字可以是正数或负数＆“这应该只出现在负数大米数字上，而不是正数大米数字上）。

import pandas as pd
import numpy as np
df1=pd.DataFrame({
        "rice": [49235.6,-49234.5,-49230.05,-49190.2,49193.1,-49140.95,-49125.3],
        "outcome": [-68.8499999999985,np.nan,np.nan,np.nan,-41.4000000000014,np.nan,np.nan,],
        "first negative from last ": [np.nan,np.nan,np.nan,np.nan,np.nan,"this",np.nan],
       })

我试图使用

df.shit()

但不确定如何提出条件。

为了更好地理解，以下 2 df 准确显示了“this”将出现在哪一行。

df2=pd.DataFrame({
        "rice": [49235.6,-49234.5,-49230.05,-49190.2,49193.1,49140.95,-49125.3],
        "outcome": [-68.8499999999985,np.nan,np.nan,np.nan,56,np.nan,np.nan,],
        "first negative from last ": [np.nan,np.nan,np.nan,np.nan,np.nan,np.nan,"this"],
       })




df3=pd.DataFrame({
        "rice": [49235.6,49234.5,-49230.05,-49190.2,-49193.1,-49140.95,-49125.3],
        "outcome": [6.8499999999985,np.nan,np.nan,np.nan,np.nan,np.nan,np.nan,],
        "first negative from last ": [np.nan,np.nan,"this",np.nan,np.nan,np.nan,np.nan],
       })
pandas

2 个回答

  1. Best Answer

    代码

    import numpy as np
cond = (df1['outcome'].bfill().isna() & df1['rice'].lt(0))
df1['first negative from last'] = (
    np.where(cond & (df1.index == cond.idxmax()), 'this', None)
)

    df1：

           rice  outcome first negative from last
0  49235.60   -68.85                     None
1 -49234.50      NaN                     None
2 -49230.05      NaN                     None
3 -49190.20      NaN                     None
4  49193.10   -41.40                     None
5 -49140.95      NaN                     this
6 -49125.30      NaN                     None
    • 1

  2. 您可以构建掩码来识别感兴趣的区域（在最后一个非 NA 和负数之后），然后使用以下命令获取第一个匹配项（如果有）idxmax：

    # mask to get values after last outcome
m1 = df1.loc[::-1, 'outcome'].isna().cummin()
# mask to get negative numbers
m2 = df1['rice'].lt(0)

s = df1.loc[m1&m2, 'rice'].isna()
if not s.empty:
    df1.loc[s.idxmax(), 'first negative from last'] = 'this'

    或者，使用shift：

    m = df1['outcome'].notna()
df1['first negative from last'] = (
    m.map({True: 'this'}).shift().mask(m[::-1].cummax())
)

    输出：

           rice  outcome first negative from last
0  49235.60   -68.85                      NaN
1 -49234.50      NaN                      NaN
2 -49230.05      NaN                      NaN
3 -49190.20      NaN                      NaN
4  49193.10   -41.40                      NaN
5 -49140.95      NaN                     this
6 -49125.30      NaN                      NaN

    中间体（第一种方法）：

           rice  outcome     m1     m2  m1&m2 first negative from last
0  49235.60   -68.85  False  False  False                      NaN
1 -49234.50      NaN  False   True  False                      NaN
2 -49230.05      NaN  False   True  False                      NaN
3 -49190.20      NaN  False   True  False                      NaN
4  49193.10   -41.40  False  False  False                      NaN
5 -49140.95      NaN   True   True   True                     this
6 -49125.30      NaN   True   True   True                      NaN

    中间体（第二种方法）：

           rice  outcome      m   map shift  cummax first negative from last
0  49235.60   -68.85   True  this  None    True                      NaN
1 -49234.50      NaN  False   NaN  this    True                      NaN
2 -49230.05      NaN  False   NaN   NaN    True                      NaN
3 -49190.20      NaN  False   NaN   NaN    True                      NaN
4  49193.10   -41.40   True  this   NaN    True                      NaN
5 -49140.95      NaN  False   NaN  this   False                     this
6 -49125.30      NaN  False   NaN   NaN   False                      NaN
    • 1

相关问题