Pandas：根据新值编辑索引值并重新分组

Question

I have my index set to 'ShiftId' which looks like this: 201912240 (the date followed by a 0 or 1 that indicates day or night shift). I have my df grouped by index values that return, as expected, something like this:

           col1 col2
201912240  NaN  23
201912241  44   75
201912250  12   NaN
201912251  46   91

I want to regroup this dataframe to take the mean of each day (ignoring NaN values), then it will look like this

           col1 col2
20191224   44   49
20191225   29   91 

But I can't get the current index values to be grouped. I have tried

    days_frame.index = days_frame.index.map(lambda x: str(x)[:-1])
    days_frame.groupby(days_frame.index).mean()

But this doesn't even change anything in the df?

Please help

Answer 1

Your solution for me working, maybe you forget assign output to variable like df here:

days_frame.index = days_frame.index.map(lambda x: str(x)[:-1])
df = days_frame.groupby(days_frame.index).mean()
print (df)
          col1  col2
20191224  44.0  49.0
20191225  29.0  91.0

Another solution rename index first and then use mean per index values:

df = days_frame.rename(lambda x: str(x)[:-1]).mean(level=0)
print (df)
          col1  col2
20191224  44.0  49.0
20191225  29.0  91.0

Or convert index to strings, remove last value and pass to groupby with aggregate mean :

df = days_frame.groupby(days_frame.index.astype(str).str[:-1]).mean()
print (df)
          col1  col2
20191224  44.0  49.0
20191225  29.0  91.0

EDIT:

If want avoid truncating all columns without A column use this solution before write to file:

df = pd.DataFrame({'A':[.41,1.5,.2,2,.3],
                   'B':['a'] * 5,
                   'C':[3,4,5,4,5],
                   'D':[1.0,3,4,5,6]})

cols = df.columns.difference(['A'])

df[cols] = df[cols].applymap(lambda x: '%.0f' % x if isinstance(x, (float, int)) else x)
print (df)

      A  B  C  D
0  0.41  a  3  1
1  1.50  a  4  3
2  0.20  a  5  4
3  2.00  a  4  5
4  0.30  a  5  6