How do I convert a numpy array into a pandas dataframe?

Question

I would like to have the 3 columns of a numpy array

px[:,:,0]
px[:,:,1]
px[:,:,0]

into a pandas Dataframe.

Should I use?

df = pd.DataFrame(px, columns=['R', 'G', 'B'])

Thank you

Hugo

May I slice through columns? df = pd.DataFrame((px[:,:,0],px[:,:,1],px[:,:,2]), columns=['R', 'G', 'B']) — Hugo
– Hugo, Commented Feb 21, 2014 at 16:01
[[[ 16 47 3] [ 14 45 1] [ 6 39 0] ..., [ 40 71 39] [ 42 73 41] [ 46 77 45]] [[ 8 41 0] [ 2 35 0] [ 23 56 11] ..., [ 35 69 32] [ 38 72 37] [ 41 75 40]] [[ 6 39 0] [ 24 60 14] [ 31 67 21] — Hugo
– Hugo, Commented Feb 21, 2014 at 16:17

Alvaro Fuentes · Accepted Answer · 2014-02-21 16:41:56Z

24

You need to reshape your array first, try this:

px2 = px.reshape((-1,3))
df = pd.DataFrame({'R':px2[:,0],'G':px2[:,1],'B':px2[:,2]})

answered Feb 21, 2014 at 16:41

Alvaro Fuentes

17.5k4 gold badges59 silver badges68 bronze badges

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

1 Answer 1