How to add numpy array elements row-wise to a pandas dataframe?

Question

I have a numpy array arr of the form:

array([[ 0.00021284, -0.04443965,  0.03926146, ...,  0.04830161,
    -0.11913304,  0.03370821],
   [ 0.01778569, -0.05192029, -0.00792321, ..., -0.01799901,
    -0.09819183,  0.06020728],
   [-0.00748426, -0.02401578,  0.01762747, ...,  0.09334017,
    -0.11837556,  0.00603597],
   [-0.03505319, -0.01932572, -0.03248611, ...,  0.00356432,
    -0.082398  ,  0.03887841],
   [-0.05111802, -0.0309066 ,  0.03542011, ..., -0.01343899,
    -0.10434885, -0.0315006 ]], dtype=float32)

Assume the shape is (5, 512)

I also have a pandas dataframe df of the form:

    Message
0   How are you?
1   What is your name?
2   What do you do?
3   What is your address?
4   Let's hang out?

I would like to attach each row in arr as an element in df by creating a new column:

    Message                Vector
0   How are you?           [ 0.00021284, -0.04443965,  0.03926146, ...,  0.04830161, -0.11913304, 0.03370821] 
1   What is your name?     [ 0.01778569, -0.05192029, -0.00792321, ..., -0.01799901, -0.09819183,  0.06020728]
2   What do you do?        [-0.00748426, -0.02401578,  0.01762747, ...,  0.09334017, -0.11837556,  0.00603597]
3   What is your address?  [-0.03505319, -0.01932572, -0.03248611, ...,  0.00356432, -0.082398,  0.03887841]
4   Let's hang out?        [-0.05111802, -0.0309066 ,  0.03542011, ..., -0.01343899, -0.10434885, -0.0315006 ]

What is an efficient way to achieve this?

Do you want the vectors in their own columns or in on big list like this? — run-out
– run-out, Commented Jan 30, 2020 at 19:21
When you say in their own columns, how would that look like? Not able to picture it, sorry. — Melsauce
– Melsauce, Commented Jan 30, 2020 at 19:21
Well your vectors column could be a column for each item in the list. Or you could have it like you do there. I'll do it both ways — run-out
– run-out, Commented Jan 30, 2020 at 19:22
Got it, so in this case like 512 columns? No, I want them to be all under one column. I know we can do pd.DataFrame(arr) and then stack it horizontally with the dataframe for the 512 column structure, but that is not what I am trying for — Melsauce
– Melsauce, Commented Jan 30, 2020 at 19:24

run-out · Accepted Answer · 2020-01-30 19:52:19Z

1

Creating an array for the problem, and convert this to a list.

a = np.array([[ 0.00021284, -0.04443965,  0.03926146,  0.04830161,
    -0.11913304,  0.03370821],
   [ 0.01778569, -0.05192029, -0.00792321, -0.01799901,
    -0.09819183,  0.06020728],
   [-0.00748426, -0.02401578,  0.01762747,  0.09334017,
    -0.11837556,  0.00603597],
   [-0.03505319, -0.01932572, -0.03248611,  0.00356432,
    -0.082398  ,  0.03887841],
   [-0.05111802, -0.0309066 ,  0.03542011, -0.01343899,
    -0.10434885, -0.0315006 ]]).tolist()

Results in:

print(a)

[[0.00021284, -0.04443965, 0.03926146, 0.04830161, -0.11913304, 0.03370821], [0.01778569, -0.05192029, -0.00792321, -0.01799901, -0.09819183, 0.06020728], [-0.00748426, -0.02401578, 0.01762747, 0.09334017, -0.11837556, 0.00603597], [-0.03505319, -0.01932572, -0.03248611, 0.00356432, -0.082398, 0.03887841], [-0.05111802, -0.0309066, 0.03542011, -0.01343899, -0.10434885, -0.0315006]]

Then add the list to the dataframe.

df = pd.DataFrame({"Message": [
"How are you?",
"What is your name?",
"What do you do?",
"What is your address?",
"Let's hang out?"]})
df['Array'] = a
print(df)

For:

                Message                                              Array
0           How are you?  [0.00021284, -0.04443965, 0.03926146, 0.048301...
1     What is your name?  [0.01778569, -0.05192029, -0.00792321, -0.0179...
2        What do you do?  [-0.00748426, -0.02401578, 0.01762747, 0.09334...
3  What is your address?  [-0.03505319, -0.01932572, -0.03248611, 0.0035...
4        Let's hang out?  [-0.05111802, -0.0309066, 0.03542011, -0.01343...

To create everything at the beginning, you can use dictionary:

df = pd.DataFrame({"Message": [
"How are you?",
"What is your name?",
"What do you do?",
"What is your address?",
"Let's hang out?"], "Array": a})

edited Jan 30, 2020 at 19:52

answered Jan 30, 2020 at 19:44

run-out

3,2041 gold badge13 silver badges27 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Melsauce Over a year ago

Thanks @run-out. The generated list has a long trail of digits. I imagine this occurs because of the original float32 datatype? Regardless, is there any way to fix it to avoid changing the representation?

YOLO Over a year ago

@run-out have you tested the solution? doesn't work for me.

run-out Over a year ago

The long trail of digits is display due to lengthy cell. The values are there.

Collectives™ on Stack Overflow

How to add numpy array elements row-wise to a pandas dataframe?

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related