Updating a row in a dataframe with values from a numpy array or list

Question

I have a dataframe, and I want to iterate change the values of some rows depending on some calculations I'm doing in a loop.

So for example: if a condition is met, Then I want to change the centers, that are the values in a row of my dataframe.

This are my centers:

       centers=[np.array([ 4.73478261,  3.10869565,  1.44782609, 0.20434783]),
       np.array([ 5.        ,  2.4       ,  3.2       ,  1.03333333]),
       np.array([ 5.135,  3.555,  1.48 ,  0.275]),
       np.array([ 5.52857143,  4.04285714,  1.47142857,  0.28571429]),
      np.array([ 5.596,  2.664,  4.052,  1.252]),
      np.array([ 6.01176471,  2.71176471,  4.94705882,  1.79411765]),
      np.array([ 6.4       ,  2.97058824,  4.55294118,  1.41176471]),
      np.array([ 6.49090909,  2.9       ,  5.37272727,  1.8       ]),
      np.array([ 6.61333333,  3.16      ,  5.56666667,  2.28666667]),
      np.array([ 7.475,  3.125,  6.3  ,  2.05 ])]

I then convert them to a dataframe

    centersDf = pd.DataFrame(centers)
    centersDf

and I would like to do something like,

    centersDf[i]=np.array[5,  1,  0  ,  2 ]

This doesn't work, but what could be the equivalent? So, I'm recalculating the centers in my loop, and I want to update my dataframe.

Assign to centersDf.iloc[i] instead of centersDf[i]. The latter refers to the i'th column. — hilberts_drinking_problem
– hilberts_drinking_problem, Commented Jul 21, 2017 at 16:03

galaxyan · Accepted Answer · 2017-07-21 16:05:14Z

1

centersDf = pd.DataFrame(centers)
centersDf.head()
   0         1         2         3       
0  4.734783  3.108696  1.447826  0.204348
1  5.000000  2.400000  3.200000  1.033333
2  5.135000  3.555000  1.480000  0.275000
3  5.528571  4.042857  1.471429  0.285714
4  5.596000  2.664000  4.052000  1.252000

centersDf.iloc[0] = np.array([5,  1,  0  ,  2 ])
centersDf.head()
   0         1         2         3       
0  5.000000  1.000000  0.000000  2.000000
1  5.000000  2.400000  3.200000  1.033333
2  5.135000  3.555000  1.480000  0.275000
3  5.528571  4.042857  1.471429  0.285714
4  5.596000  2.664000  4.052000  1.252000

answered Jul 21, 2017 at 16:05

galaxyan

6,1593 gold badges23 silver badges44 bronze badges

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

Comments

piRSquared · Accepted Answer · 2017-07-21 16:23:07Z

When you pass a scalar value to the __getitem__ method (that is using the []) pandas mathches up against column names. So centersDf[0] is the 0th column. You get an error because your trying to assign an array of length 4 to a column of length 10 and that makes no sense.

If you want to be able to assign by column name, create your dataframe as it's transpose

centersDf = pd.DataFrame(centers).T

Then

centersDf[0] = [5,  1,  0 ,  2]

Works fine

centersDf

   0         1      2         3      4         5         6         7         8      9
0  5  5.000000  5.135  5.528571  5.596  6.011765  6.400000  6.490909  6.613333  7.475
1  1  2.400000  3.555  4.042857  2.664  2.711765  2.970588  2.900000  3.160000  3.125
2  0  3.200000  1.480  1.471429  4.052  4.947059  4.552941  5.372727  5.566667  6.300
3  2  1.033333  0.275  0.285714  1.252  1.794118  1.411765  1.800000  2.286667  2.050

Otherwise, just use loc as has been suggested already.

Collectives™ on Stack Overflow

Updating a row in a dataframe with values from a numpy array or list

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related