如何给行添加值？

我创建在数据帧中的一个列和分配值0。
写逻辑来更新这些现在列值，但不反射的。

输入：

>>> parafix_df = main_df[["line_width", "para_num", "bbox" ]] 
>>> parafix_df 
    line_width para_num        bbox 
0  238.546  NaN (50.0, 579.3, 288.546, 598.022) 
1   318  1 (64.0, 564.9, 382.0, 583.622) 
2   332  2 (50.0, 550.5, 382.0, 569.222) 
3   332  2 (50.0, 536.1, 382.0, 554.822) 
4  328.977  2 (50.0, 521.7, 378.977, 540.422) 
5   318  3 (64.0, 507.3, 382.0, 526.022) 
6   332  3 (50.0, 492.9, 382.0, 511.622) 
7   332  3 (50.0, 478.5, 382.0, 497.222) 
8   332  3 (50.0, 464.1, 382.0, 482.822) 
9   332  3 (50.0, 449.7, 382.0, 468.422) 
10  59.04  3 (50.0, 435.3, 109.04, 454.022) 
11 304.007  4 (64.0, 420.9, 368.007, 439.622) 
12  318  5 (64.0, 406.5, 382.0, 425.222) 
13  332  5 (50.0, 392.1, 382.0, 410.822) 
14  332  5 (50.0, 377.7, 382.0, 396.422) 
15  332  5 (50.0, 363.3, 382.0, 382.022) 
16  43.252  5 (50.0, 348.9, 93.252, 367.622) 
17  318  6 (64.0, 334.5, 382.0, 353.222) 
18  332  6 (50.0, 320.1, 382.0, 338.822) 
19  332  6 (50.0, 305.7, 382.0, 324.422) 
20  332  6 (50.0, 291.3, 382.0, 310.022) 
21  332  6 (50.0, 276.9, 382.0, 295.622) 
22  317.02  6 (50.0, 262.5, 367.02, 281.222) 
23  318  7 (64.0, 248.1, 382.0, 266.822) 
24  332  7 (50.0, 233.7, 382.0, 252.422) 
25  47.014  7 (50.0, 219.3, 97.014, 238.022) 
26  318  8 (64.0, 204.9, 382.0, 223.622) 
27 316.723  8 (50.0, 190.5, 366.723, 209.222) 
28  318  9 (64.0, 176.1, 382.0, 194.822) 
29 326.766  9 (50.0, 161.7, 376.766, 180.422) 
30  318  10 (64.0, 147.3, 382.0, 166.022) 
31  332  10 (50.0, 132.9, 382.0, 151.622) 
32  332  10 (50.0, 118.5, 382.0, 137.222) 
33 305.393  11 (64.0, 104.1, 369.393, 122.822) 
34  318  12  (64.0, 89.7, 382.0, 108.422) 
35  318  13  (64.0, 75.3, 382.0, 94.022) 
36 319.165  13 (50.0, 60.9, 369.165, 79.622) 
37 308.165  14 (64.0, 46.5, 372.165, 65.222) 
38  318  15  (64.0, 32.1, 382.0, 50.822) 
39 329.153  15 (50.0, 17.7, 379.153, 36.422) 
40  318  16  (64.0, 3.3, 382.0, 22.022) 
41 324.335  16 (50.0, -11.1, 374.335, 7.622)

代码：

       bbox new_para_num 
0 (50.0, 579.3, 288.546, 598.022)    0 
1  (64.0, 564.9, 382.0, 583.622)    0 
2  (50.0, 550.5, 382.0, 569.222)    0 
3  (50.0, 536.1, 382.0, 554.822)    0 
4 (50.0, 521.7, 378.977, 540.422)    0 
5  (64.0, 507.3, 382.0, 526.022)    0 
6  (50.0, 492.9, 382.0, 511.622)    0 
7  (50.0, 478.5, 382.0, 497.222)    0 
8  (50.0, 464.1, 382.0, 482.822)    0 
9  (50.0, 449.7, 382.0, 468.422)    0 
10 (50.0, 435.3, 109.04, 454.022)    0 
11 (64.0, 420.9, 368.007, 439.622)    0 
12 (64.0, 406.5, 382.0, 425.222)    0 
13 (50.0, 392.1, 382.0, 410.822)    0 
14 (50.0, 377.7, 382.0, 396.422)    0 
15 (50.0, 363.3, 382.0, 382.022)    0 
16 (50.0, 348.9, 93.252, 367.622)    0 
17 (64.0, 334.5, 382.0, 353.222)    0 
18 (50.0, 320.1, 382.0, 338.822)    0 
19 (50.0, 305.7, 382.0, 324.422)    0 
20 (50.0, 291.3, 382.0, 310.022)    0 
21 (50.0, 276.9, 382.0, 295.622)    0 
22 (50.0, 262.5, 367.02, 281.222)    0 
23 (64.0, 248.1, 382.0, 266.822)    0 
24 (50.0, 233.7, 382.0, 252.422)    0 
25 (50.0, 219.3, 97.014, 238.022)    0 
26 (64.0, 204.9, 382.0, 223.622)    0 
27 (50.0, 190.5, 366.723, 209.222)    0 
28 (64.0, 176.1, 382.0, 194.822)    0 
29 (50.0, 161.7, 376.766, 180.422)    0 
30 (64.0, 147.3, 382.0, 166.022)    0 
31 (50.0, 132.9, 382.0, 151.622)    0 
32 (50.0, 118.5, 382.0, 137.222)    0 
33 (64.0, 104.1, 369.393, 122.822)    0 
34  (64.0, 89.7, 382.0, 108.422)    0 
35  (64.0, 75.3, 382.0, 94.022)    0 
36 (50.0, 60.9, 369.165, 79.622)    0 
37 (64.0, 46.5, 372.165, 65.222)    0 
38  (64.0, 32.1, 382.0, 50.822)    0 
39 (50.0, 17.7, 379.153, 36.422)    0 
40  (64.0, 3.3, 382.0, 22.022)    0 
41 (50.0, -11.1, 374.335, 7.622)    0

：上述代码

parafix_df = main_df[["line_text", "line_width", "para_num", "bbox" ]] 
parafix_df["new_para_num"] = 0 

max_width = parafix_df['line_width'].max() 
bbox_max_width = parafix_df.loc[selected['line_width'] == max_width].iloc[0]["bbox"] 

previous = None 
para1 = 1 
for current, next in izip(parafix_df.iterrows(), parafix_df.iloc[1:].iterrows()): 
    if previous==None: 
     current[1]["new_para_num"] = para1 
    else: 
     bbox_current = current[1]["bbox"] 
     bbox_next = next[1]["bbox"] 
     bbox_previous = previous[1]["bbox"] 
     if bbox_current[0]>bbox_max_width[0]: 
      para1 += 1 
      print "para1:", para1 
     current[1]["new_para_num"] = para1 

    previous = current

输出

但我希望新的对数值：

para1: 2 para1: 3 para1: 4 para1: 5 para1: 6 para1: 7 para1: 8 para1: 9 para1: 10 para1: 11 para1: 12 para1: 13 para1: 14 para1: 15 para1: 16

你能帮助我吗？

以下是我最后的工作代码：

parafix_df = main_df[["line_text", "line_width", "para_num", "bbox" ]] parafix_df["new_para_num"] = 0 max_width = parafix_df['line_width'].max() bbox_max_width = parafix_df.loc[selected['line_width'] == max_width].iloc[0]["bbox"] para1 = 1 for indx, current in enumerate(parafix_df.iterrows(), start=0): if indx!=0: bbox_current = current[1]["bbox"] if bbox_current[0]>bbox_max_width[0]: para1 += 1 parafix_df.iloc[indx, 4] = para1

我们可以优化吗？

来源

2016-05-26 Vivek Sable

在你的代码中总是只访问第二行（索引：'1'） - 'current [1]'，'next [1]'。但我想你不需要通过你的DF循环来实现你的目标 - 应该可以用“熊猫”的方式做到这一点 – MaxU

是的，我是熊猫新手。你可以再次检查我的代码，我更新了代码。 –

UPDATE：

IIUC，你可以这样来做：

df.new_para_num = 1 

In [210]: df.loc[df.line_width == df.line_width.max(), 'new_para_num'].cumsum() + 1 
Out[210]: 
2  2 
3  3 
6  4 
7  5 
8  6 
9  7 
13  8 
14  9 
15 10 
18 11 
19 12 
20 13 
21 14 
24 15 
31 16 
32 17 
Name: new_para_num, dtype: int64

如果你想在你原来的DF更新 new_para_num列有条件

：

In [223]: df.new_para_num = 1 

In [224]: selected = df.loc[df.line_width == df.line_width.max()].copy() 

In [226]: selected.new_para_num = selected.new_para_num.cumsum() + 1 

In [227]: selected 
Out[227]: 
    line_width para_num       bbox new_para_num 
2  332.0  2.0 [50.0, 550.5, 382.0, 569.222]    2 
3  332.0  2.0 [50.0, 536.1, 382.0, 554.822]    3 
6  332.0  3.0 [50.0, 492.9, 382.0, 511.622]    4 
7  332.0  3.0 [50.0, 478.5, 382.0, 497.222]    5 
8  332.0  3.0 [50.0, 464.1, 382.0, 482.822]    6 
9  332.0  3.0 [50.0, 449.7, 382.0, 468.422]    7 
13  332.0  5.0 [50.0, 392.1, 382.0, 410.822]    8 
14  332.0  5.0 [50.0, 377.7, 382.0, 396.422]    9 
15  332.0  5.0 [50.0, 363.3, 382.0, 382.022]   10 
18  332.0  6.0 [50.0, 320.1, 382.0, 338.822]   11 
19  332.0  6.0 [50.0, 305.7, 382.0, 324.422]   12 
20  332.0  6.0 [50.0, 291.3, 382.0, 310.022]   13 
21  332.0  6.0 [50.0, 276.9, 382.0, 295.622]   14 
24  332.0  7.0 [50.0, 233.7, 382.0, 252.422]   15 
31  332.0  10.0 [50.0, 132.9, 382.0, 151.622]   16 
32  332.0  10.0 [50.0, 118.5, 382.0, 137.222]   17 

In [228]: df.loc[df.line_width == df.line_width.max(), 'new_para_num'] = selected 

In [229]: df 
Out[229]: 
    line_width para_num        bbox new_para_num 
0  238.546  NaN [50.0, 579.3, 288.546, 598.022]    1 
1  318.000  1.0 [64.0, 564.9, 382.0, 583.622]    1 
2  332.000  2.0 [50.0, 550.5, 382.0, 569.222]    2 
3  332.000  2.0 [50.0, 536.1, 382.0, 554.822]    3 
4  328.977  2.0 [50.0, 521.7, 378.977, 540.422]    1 
5  318.000  3.0 [64.0, 507.3, 382.0, 526.022]    1 
6  332.000  3.0 [50.0, 492.9, 382.0, 511.622]    4 
7  332.000  3.0 [50.0, 478.5, 382.0, 497.222]    5 
8  332.000  3.0 [50.0, 464.1, 382.0, 482.822]    6 
9  332.000  3.0 [50.0, 449.7, 382.0, 468.422]    7 
10  59.040  3.0 [50.0, 435.3, 109.04, 454.022]    1 
11  304.007  4.0 [64.0, 420.9, 368.007, 439.622]    1 
12  318.000  5.0 [64.0, 406.5, 382.0, 425.222]    1 
13  332.000  5.0 [50.0, 392.1, 382.0, 410.822]    8 
14  332.000  5.0 [50.0, 377.7, 382.0, 396.422]    9 
15  332.000  5.0 [50.0, 363.3, 382.0, 382.022]   10 
16  43.252  5.0 [50.0, 348.9, 93.252, 367.622]    1 
17  318.000  6.0 [64.0, 334.5, 382.0, 353.222]    1 
18  332.000  6.0 [50.0, 320.1, 382.0, 338.822]   11 
19  332.000  6.0 [50.0, 305.7, 382.0, 324.422]   12 
20  332.000  6.0 [50.0, 291.3, 382.0, 310.022]   13 
21  332.000  6.0 [50.0, 276.9, 382.0, 295.622]   14 
22  317.020  6.0 [50.0, 262.5, 367.02, 281.222]    1 
23  318.000  7.0 [64.0, 248.1, 382.0, 266.822]    1 
24  332.000  7.0 [50.0, 233.7, 382.0, 252.422]   15 
25  47.014  7.0 [50.0, 219.3, 97.014, 238.022]    1 
26  318.000  8.0 [64.0, 204.9, 382.0, 223.622]    1 
27  316.723  8.0 [50.0, 190.5, 366.723, 209.222]    1 
28  318.000  9.0 [64.0, 176.1, 382.0, 194.822]    1 
29  326.766  9.0 [50.0, 161.7, 376.766, 180.422]    1 
30  318.000  10.0 [64.0, 147.3, 382.0, 166.022]    1 
31  332.000  10.0 [50.0, 132.9, 382.0, 151.622]   16 
32  332.000  10.0 [50.0, 118.5, 382.0, 137.222]   17 
33  305.393  11.0 [64.0, 104.1, 369.393, 122.822]    1 
34  318.000  12.0  [64.0, 89.7, 382.0, 108.422]    1 
35  318.000  13.0  [64.0, 75.3, 382.0, 94.022]    1 
36  319.165  13.0 [50.0, 60.9, 369.165, 79.622]    1 
37  308.165  14.0 [64.0, 46.5, 372.165, 65.222]    1 
38  318.000  15.0  [64.0, 32.1, 382.0, 50.822]    1 
39  329.153  15.0 [50.0, 17.7, 379.153, 36.422]    1 
40  318.000  16.0  [64.0, 3.3, 382.0, 22.022]    1 
41  324.335  16.0 [50.0, -11.1, 374.335, 7.622]    1

PS，但我仍然不确定我是否正确理解你的目标

OLD答案：

您可以使用shift功能，以获得一个和下一个行：

df.shift(-1) # df will be shifted one row backwards (will show `next` row) 

df.shift(1) # df will be shifted one row forwards (will show `prev` row)

例子：

In [142]: df 
Out[142]: 
    a b c 
0 8 3 0 
1 8 3 4 
2 9 4 1 
3 2 1 8 
4 5 6 3 

In [147]: df['prev_a'] = df.a.shift(1) 

In [148]: df['next_a'] = df.a.shift(-1) 

In [149]: df 
Out[149]: 
    a b c prev_a next_a 
0 8 3 0  NaN  8.0 
1 8 3 4  8.0  9.0 
2 9 4 1  8.0  2.0 
3 2 1 8  9.0  5.0 
4 5 6 3  2.0  NaN

来源

2016-05-26 11:32:59 MaxU

谢谢你+10，但是我无法在'new_para_num'列中反映出变化 –

@VivekSable，你可以发布样本_input_和想要的_output_数据集（CSV/dict/JSON/Python代码格式5-7行__as text__，所以在编码时可以使用它）并描述你想如何处理输入数据以获得输出数据集？ [如何创建一个最小，完整和可验证的示例]（http：// stackoverflow。com/help/mcve） – MaxU

我想根据逻辑值给新列赋值。（逻辑我将更新） –

如何给行添加值？

回答

相关问题