我尝试使用sort_index重新排序列，但它的行为并不像我希望的那样。
排序后，重新排序如下：
>>> list(df)
['blocks_written_1',
'blocks_written_10',
'blocks_written_11',
'blocks_written_12',
'blocks_written_2',
'blocks_written_3',
'blocks_written_4',
'blocks_written_5',
'blocks_written_6',
'blocks_written_7',
'blocks_written_8',
'blocks_written_9']
但我正在寻找以下内容：
blocks_written_1,
blocks_written_2,
..,
blocks_written_9,
blocks_written_10,
blocks_written_11,
blocks_written_12
感谢任何帮助！
答案 0 :(得分：0)
# Create sample data.
n = ['1', '10', '11', '12', '2', '3', '4', '5', '6', '7', '8', '9']
initial_cols = ['blocks_written_{}'.format(i) for i in n]
np.random.seed(0)
df = pd.DataFrame(np.random.randn(5, len(initial_cols)), columns=initial_cols)
# Split column names on underscore and take integer of last item.
cols = [(col, int(col.split('_')[-1])) for col in df]
# Sort column name, integer tuple pairs on integer value.
cols.sort(key=lambda x: x[1])
# Rearrange dataframe based on new column order.
df = df[[col[0] for col in cols]]
>>> df
blocks_written_1 blocks_written_2 blocks_written_3 blocks_written_4 \
0 1.764052 1.867558 -0.977278 0.950088
1 0.761038 1.494079 -0.205158 0.313068
2 2.269755 1.532779 1.469359 0.154947
3 1.230291 -1.048553 -1.420018 -1.706270
4 -1.613898 -0.510805 -1.180632 -0.028182
blocks_written_5 blocks_written_6 blocks_written_7 blocks_written_8 \
0 -0.151357 -0.103219 0.410599 0.144044
1 -0.854096 -2.552990 0.653619 0.864436
2 0.378163 -0.887786 -1.980796 -0.347912
3 1.950775 -0.509652 -0.438074 -1.252795
4 0.428332 0.066517 0.302472 -0.634322
blocks_written_9 blocks_written_10 blocks_written_11 blocks_written_12
0 1.454274 0.400157 0.978738 2.240893
1 -0.742165 0.121675 0.443863 0.333674
2 0.156349 -1.454366 0.045759 -0.187184
3 0.777490 1.202380 -0.387327 -0.302303
4 -0.362741 -0.212740 -0.895467 0.386902