重新排序行数据

Question

Based on this question Can't Re-Order Columns Data , how to do it for the dataframe row? because I have a problem, not in the column order but the row order.

my data looks like this:

    B1  B2
B1  1   1
B10 1   1
B11 1   1
B12 1   1
B2  1   1
B20 1   1
B21 1   1
B22 1   1
B3  1   1
B30 1   1
B31 1   1

my expected result:

    B1  B2
 B1 1   1
 B2 1   1
 B3 1   1
B10 1   1
B11 1   1
B12 1   1
B20 1   1
B21 1   1
B22 1   1
B30 1   1
B31 1   1

Answer 1

you can use natsort

import natsort as ns
df.reindex(ns.natsorted(df.index))

---

     B1  B2
B1    1   1
B2    1   1
B3    1   1
B10   1   1
B11   1   1
B12   1   1
B20   1   1
B21   1   1
B22   1   1
B30   1   1
B31   1   1

Or:

i=df.index.to_series().str.extract('(\d+)',expand=False).astype(float).sort_values().index
df.reindex(i)

Answer 2

You can use parameter key in sorted function and pass output to DataFrame.reindex :

df = df.reindex(sorted(df.index, key=lambda x: float(x[1:])))
print (df)
     B1  B2
B1    1   1
B2    1   1
B3    1   1
B10   1   1
B11   1   1
B12   1   1
B20   1   1
B21   1   1
B22   1   1
B30   1   1
B31   1   1

Alternative is Series.str.extract numeric, convert to floats and get positions of sorted values by Index.argsort , last change order by DataFrame.iloc :

df = df.iloc[df.index.str.extract('(\d+)', expand=False).astype(float).argsort()]
print (df)
     B1  B2
B1    1   1
B2    1   1
B3    1   1
B10   1   1
B11   1   1
B12   1   1
B20   1   1
B21   1   1
B22   1   1
B30   1   1
B31   1   1