Pandas 基於行的最大值和最小值舍入 function

Question

使用以下函數round_up和round_down我可以返回包含四舍五入 output 的列，如下所示：

            Open    High    Low     Close  round_up round_down
date                        
2020-03-19  86.55   86.88   85.83   85.83   87.0    85.0
2020-03-20  85.94   86.30   85.20   85.23   87.0    85.0
2020-03-21  85.06   85.18   84.13   84.23   86.0    84.0
2020-03-22  84.37   84.72   83.62   84.72   85.0    83.0
2020-03-23  84.83   85.70   84.41   85.70   86.0    84.0
2020-03-24  85.87   86.79   85.58   86.73   87.0    85.0
2020-03-25  86.90   86.90   86.15   86.87   87.0    86.0
2020-03-26  87.18   88.92   87.15   88.65   89.0    87.0
2020-03-27  88.69   88.95   87.90   88.40   89.0    87.0
2020-03-28  88.31   88.50   87.69   87.92   89.0    87.0
2020-03-29  87.84   87.84   86.04   86.31   88.0    86.0
2020-03-30  86.27   86.38   85.57   86.30   87.0    85.0
2020-03-31  86.72   87.23   86.24   86.90   88.0    86.0
2020-04-01  86.83   87.48   86.65   86.65   88.0    86.0
2020-04-02  86.53   86.75   85.32   86.13   87.0    85.0

如何修改這些函數以返回所需的 output ，其中round up和round down入到最接近的5 ：

            Open    High    Low     Close  round_up round_down
date                        
2020-03-19  86.55   86.88   85.83   85.83   90.0    85.0
2020-03-20  85.94   86.30   85.20   85.23   90.0    85.0
2020-03-21  85.06   85.18   84.13   84.23   90.0    80.0
2020-03-22  84.37   84.72   83.62   84.72   85.0    80.0

重現上述數據的代碼

def genMockDataFrame(days,startPrice,colName,startDate,seed=None,round_to=4,scale=0.0018): 
   
    periods = days*24
    np.random.seed(seed)
    steps = np.random.normal(loc=0, scale=scale, size=periods)
    steps[0]=0
    P = startPrice+np.cumsum(steps)
    P = [round(i,round_to) for i in P]

    fxDF = pd.DataFrame({ 
        'ticker':np.repeat( [colName], periods ),
        'date':np.tile( pd.date_range(startDate, periods=periods, freq='H'), 1 ),
        'price':(P)})
    fxDF.index = pd.to_datetime(fxDF.date)
    fxDF = fxDF.price.resample('D').ohlc()
    fxDF.columns = [i.title() for i in fxDF.columns]
    return fxDF

ticker = 'XXX_USD'


df = genMockDataFrame(15,86.55,ticker,'19/3/2020',seed=157,round_to=2,scale=0.18)

  

def round_up(x, decimals=0):
    n = x.max()
    multiplier = 10 ** decimals
    return math.ceil(n * multiplier) / multiplier


def round_down(x, decimals=0):
    n = x.min()
    multiplier = 10 ** decimals
    return math.floor(n * multiplier) / multiplier

df["round_up"] = df.apply(round_up,decimals=0,axis=1)
df["round_down"] = df.apply(round_down,decimals=0,axis=1)

Answer 1

嘗試：

def round_up(df, base=5):
    return df.max(axis=1) + (base - df.max(axis=1)) % base

def round_down(df, base=5):
    return df.min(axis=1) - (df.min(axis=1) % base)

cols = ['Open', 'High', 'Low', 'Close']
df['round_up'] = round_up(df[cols])
df['round_down'] = round_down(df[cols])

Output：

>>> df
             Open   High    Low  Close  round_up  round_down
date                                                        
2020-03-19  86.55  86.88  85.83  85.83      90.0        85.0
2020-03-20  85.94  86.30  85.20  85.23      90.0        85.0
2020-03-21  85.06  85.18  84.13  84.23      90.0        80.0
2020-03-22  84.37  84.72  83.62  84.72      85.0        80.0
2020-03-23  84.83  85.70  84.41  85.70      90.0        80.0
2020-03-24  85.87  86.79  85.58  86.73      90.0        85.0
2020-03-25  86.90  86.90  86.15  86.87      90.0        85.0
2020-03-26  87.18  88.92  87.15  88.65      90.0        85.0
2020-03-27  88.69  88.95  87.90  88.40      90.0        85.0
2020-03-28  88.31  88.50  87.69  87.92      90.0        85.0
2020-03-29  87.84  87.84  86.04  86.31      90.0        85.0
2020-03-30  86.27  86.38  85.57  86.30      90.0        85.0
2020-03-31  86.72  87.23  86.24  86.90      90.0        85.0
2020-04-01  86.83  87.48  86.65  86.65      90.0        85.0
2020-04-02  86.53  86.75  85.32  86.13      90.0        85.0