I would like to know how to add multiple constant values of different lengths into a dataframe column. I know that we can add a single constant value (for example: 5) to a data frame column 'A' like this:
df['A'] = 5
But I want to have the dataframe something like the table below. As you can see, I need three 5s, two 10s, six 30s and one 100s. How can you do that for maybe 10000 rows with a set number of values (not random) each having a user defined frequency.
index | A |
---|---|
1 | 5 |
2 | 5 |
3 | 5 |
4 | 10 |
5 | 10 |
6 | 30 |
7 | 30 |
8 | 30 |
9 | 30 |
10 | 30 |
11 | 30 |
12 | 100 |
CodePudding user response:
IIUC you could just use:
df['b'] = np.repeat([5, 5, 5, 10, 10, 30, 30, 30, 30, 30, 30, 100], np.ceil(len(df) / 12))[:len(df)]
Or:
df['b'] = np.repeat([*[5] * 3, *[10] * 2, *[30] * 6, 100], np.ceil(len(df) / 12))[:len(df)]
CodePudding user response:
You can use numpy.repeat
with the DataFrame constructor:
vals = [5,10,30,100]
reps = [3,2,6,1]
df = pd.DataFrame({'A': np.repeat(vals, reps)})
df.index =1
output:
A
1 5
2 5
3 5
4 10
5 10
6 30
7 30
8 30
9 30
10 30
11 30
12 100
CodePudding user response:
Try this:
import itertools
value = [5,10,30,100]
repeat = [3,2,6,1]
lst = [([v]*r) for v,r in zip(value,repeat)]
merged = list(itertools.chain(*lst))
df = pd.DataFrame({'A':merged})
df
Output:
A
0 5
1 5
2 5
3 10
4 10
5 30
6 30
7 30
8 30
9 30
10 30
11 100
CodePudding user response:
You can specify values and repeats in dictionary and use:
d = {3:5, 2:10, 6:30, 1:100}
df = pd.DataFrame({'A': [x for k, v in d.items() for x in [v] * k]})
print (df)
A
0 5
1 5
2 5
3 10
4 10
5 30
6 30
7 30
8 30
9 30
10 30
11 100
Solution with dictionary and numpy.repeat
:
df = pd.DataFrame({'A': np.repeat(list(d.values()), list(d.keys()))})
print (df)
A
0 5
1 5
2 5
3 10
4 10
5 30
6 30
7 30
8 30
9 30
10 30
11 100