从以“#”分隔的行值中提取最小值和最大值

Question

我有一个包含列的数据框，

Col_1    Col_2

'0'         '-33#90#'
'-1#65#'       '0'
'90'        '-22#-44#90#250'

我想获取各个列值的最小值和最大值。我不知道如何获得它。

我的输出应该是这样的，

Col_1_min  col_1_max   col_2_min   col_2_max
0           null         -33          90
-1           65           0           null
90          null         -44          250

Answer 1

如果值相同，您可以使用 Series.str.extractall for get negative and positive integers, reshape by Series.unstack and convert to floats for numeric. Then get minimal and maximum values with Series.where 按列表中的列循环查找错误值：

cols = ['Col_1', 'Col_2']
for c in cols:
    df1 = df[c].str.extractall('([-]?\d+)')[0].unstack().astype(float)
    min1 = df1.min(axis=1)
    max1 = df1.max(axis=1)

    df[f'{c}_min'] = min1
    df[f'{c}_max'] = max1.mask(max1==min1)
print (df)
      Col_1             Col_2  Col_1_min  Col_1_max  Col_2_min  Col_2_max
0       '0'         '-33#90#'          0        NaN        -33       90.0
1  '-1#65#'               '0'         -1       65.0          0        NaN
2      '90'  '-22#-44#90#250'         90        NaN        -44      250.0

如果需要删除原始列：

cols = ['Col_1', 'Col_2']
for c in cols:
    df1 = df.pop(c).str.extractall('([-]?\d+)')[0].unstack().astype(float)
    min1 = df1.min(axis=1)
    max1 = df1.max(axis=1)
    df[f'{c}_min'] = min1
    df[f'{c}_max'] = max1.mask(max1==min1)
print (df)
   Col_1_min  Col_1_max  Col_2_min  Col_2_max
0          0        NaN        -33       90.0
1         -1       65.0          0        NaN
2         90        NaN        -44      250.0

编辑：

另一种拆分解决方案：

cols = ['Col_1', 'Col_2']
for c in cols:
    df1 = df.pop(c).str.strip("'").str.split('#', expand=True)
    df1 = df1.apply(pd.to_numeric, errors='coerce')
    min1 = df1.min(axis=1)
    max1 = df1.max(axis=1)
    df[f'{c}_min'] = min1
    df[f'{c}_max'] = max1.mask(max1==min1)
print (df)
   Col_1_min  Col_1_max  Col_2_min  Col_2_max
0        0.0        0.0      -33.0        NaN
1       -1.0        NaN        0.0        0.0
2       90.0       90.0      -44.0        NaN

从以“#”分隔的行值中提取最小值和最大值

Extract Min and Max Value from the row value delimited by "#"

python

string

max

min

pandas