I have the following dataframe :
df = pd.DataFrame(
{'a':[np.array([1,2,3]),np.array([4,5,6])]
}
)
i want to have a numpy array like following extracted from the a column:
array([[1., 2., 3.],
[4., 5., 6.]])
i can do it like below:
for i, row in df.iterrows():
t_arr[i,:] = row["a"]
but i am looking for a solution without for loop. .values and to_numpy() don't return the desired results
You could do this:
np.vstack(df['a'])
which returns:
array([[1, 2, 3],
[4, 5, 6]])
Use np.stack to create a 2d numpy array.
Either from the series object of your a column:
np.stack(df['a'].values)
Or from the whole DataFrame:
np.stack(df.values.flatten())
Edit: You don't even need to call .values. You can just call np.stack on the series object.
np.stack iterates through the argument (such as a Series) and creates a 2d array from each. Skipping the values doesn't save time. Leaving it in probably makes the action clearer to humans.Using tolist can work, followed by np.array will correctly return a (2,3) numpy array
np.array(df["a"].values.tolist())
returns
array([[1, 2, 3],
[4, 5, 6]])
values. It's a useful starting point for creating an answer. I expect it is an object dtype array with 2 array elements. In general they might differ in shape, but if they do match, they can then be joined into one array as demonstrated in the answers.