ProfileReport(df) 将计算的间隔显示为 MISSING、REJECTED、UNSUPPORTED

Question

在数据框中，我将 m/d/yyyy 格式的字符串转换为日期时间，然后减去它们以获得间隔。出于某种原因，这些计算字段在 pandas.ProfileReport() 中被描述为缺失、拒绝、不支持，我无法在 .agg 计算中使用它们

df['RequestedStartDT']  = pd.to_datetime(df['Requested Start Date'])
df['RequestedEndDT']    = pd.to_datetime(df['Requested End Date'])

df['TargetedStartDT']   = pd.to_datetime(df['Targeted Start Date'])
df['TargetedEndDT']     = pd.to_datetime(df['Targeted End Date'])

df['ActualStartDT']     = pd.to_datetime(df['Actual Start Date'])
df['ActualEndDT']       = pd.to_datetime(df['Actual End Date'])

df['RequestedInterval']       = df['RequestedEndDT']    - df['RequestedStartDT']
df['TargetedInterval']        = df['TargetedStartDT']   - df['TargetedEndDT']
df['ActualInterval']          = df['ActualEndDT']       - df['ActualStartDT']

对于计算和聚合这些间隔的更好方法有什么建议吗？

这是数据示例

Answer 1

原来我需要显式转换字段

df['RequestedInterval']       =df['RequestedInterval']/np.timedelta64(1,'D')
df['TargetedInterval']        =df['TargetedInterval']/np.timedelta64(1,'D')
df['ActualInterval']          =df['ActualInterval']/np.timedelta64(1,'D')

ProfileReport(df) 将计算的间隔显示为 MISSING、REJECTED、UNSUPPORTED

ProfileReport(df) shows calculated intervals as MISSING, REJECTED, UNSUPPORTED

python

aggregate

date-arithmetic

pandas