python - 如何向 Pandas 中的每一行广播统一的随机数？

Question

我在这个外观中有一个数据框：

    0:00   0:30   1:00   1:30   2:00   2:30   3:00   3:30   4:00   4:30
0 -1.149 -2.105 -2.036 -1.854 -1.807 -0.728 -1.181 -1.801 -0.665 -0.076
1 -1.136 -1.559 -1.332 -1.184 -0.097 -1.215 -0.386 -0.703 -1.176 -0.557
2 -1.147 -1.336 -1.326 -1.568 -1.344 -1.305 -1.398 -1.175 -1.561 -0.439
3 -1.323 -1.337 -1.326 -1.556 -1.348 -1.656 -1.046 -2.715 -1.810 -0.815
4 -1.175 -1.689 -1.505 -1.232 -0.115 -1.259 -0.406 -0.736 -1.652 -0.085
5 -1.243 -1.306 -1.355 -1.539 -1.353 -0.443 -0.853 -1.275 -0.939 -0.679
6 -1.277 -1.728 -1.453 -1.351 -1.271 -0.090 -1.503 -0.984 -0.679 -0.245
7 -1.281 -1.402 -1.491 -1.392 -1.485 -0.167 -1.414 -0.800 -0.635 -1.354
8 -2.079 -1.767 -1.453 -1.548 -1.446 -0.699 -1.235 -1.678 -0.698 -1.037
9 -1.190 -2.238 -1.988 -1.866 -0.932 -1.671 -0.858 -1.275 -2.388 -0.198

我想添加几个新行。新行应该从现有行中随机选择（选择整行），并稍微改变一下。我想对每个新行应用统一的随机数。

我尝试的是：

d_jlist = pd.read_csv('127case.csv', sep=',')
d_jlist =  d_jlist.iloc[0:10,0:10]

d_jlist4 = d_jlist
d_jlist3 = pd.DataFrame()

a = np.random.choice(range(2,8),size = 5 )     # Randomly select 5 rows from existing rows

for i in a:
    b = np.random.uniform(-1,1)
    print(b)

    d_jlist3 = d_jlist3.append(d_jlist4.iloc[i] ) + b  #broadcast the random number to every element in this row
   #d_jlist3 = d_jlist3.append(d_jlist4.iloc[i] + b )  # If I try in this way, why it will also change the existing rows?

print((d_jlist3))

d_jlist4 = d_jlist4.append(d_jlist3)          #Add the new rows to the existing ones

生成的统一随机数和新行如下所示：

0.28761993446482825
-0.15132898721462507
0.8753189320820596
-0.05574300352910355
-0.7961990667560808

       0:00      0:30      1:00  ...      3:30      4:00      4:30
6 -1.117332 -1.568332 -1.293332  ... -0.824332 -0.519332 -0.085332
6 -1.404952 -1.855952 -1.580952  ... -1.111952 -0.806952 -0.372952
7 -1.257623 -1.378623 -1.467623  ... -0.776623 -0.611623 -1.330623
5 -2.094942 -2.157942 -2.206942  ... -2.126942 -1.790942 -1.530942
4 -1.971199 -2.485199 -2.301199  ... -1.532199 -2.448199 -0.881199

我认为关系应该是旧行 + 随机数 = 新行。

但我检查了结果，它不能满足。我想知道实现这个？

提前致谢！

标签： pythonpython-3.xpandas

Answer 1

这是一种方法（假设我了解您要执行的操作）：

NUM_NEW_ROWS = 2

random_rows = df.sample(NUM_NEW_ROWS)
random_rows = random_rows.add(np.random.rand(NUM_NEW_ROWS), axis = "rows")
pd.concat([df, random_rows], axis="rows")

这是具有三列和值 1-9 的简单数据框的结果：

           a         b         c
0  1.000000  4.000000  7.000000
1  2.000000  5.000000  8.000000
2  3.000000  6.000000  9.000000
0  1.415359  4.415359  7.415359
1  2.512821  5.512821  8.512821