CODEWITHSUNDEEP
pythonpandas
Gemini
Gemini

Reputation: 475

Get Row with Closest Value

in pandas I want to introduce a new column 'Neighbor' in my dataframe which saves (e.g.) the index or the value of another row, for which a certain column B is most close to the value of the current row. E.g.

   A   Neighbor
0  5   1
1  4   0
2  1   3
3  2   2
4  10  0

Do you know a pandas function doing that? As my dataset is very big apply unfortunately is too slow.

Thanks!

Upvotes: 0

Views: 496

Answers (2)

boethius
boethius

Reputation: 438

Something like this might do the trick:

df['myindex']  = df.index
# result = df.sort(['A'])
result = df.sort_values(by='A')
result['NeighborIndex'] = result['myindex'].shift(-1)

Upvotes: 1

BENY
BENY

Reputation: 323376

IDK, whether you check my update answer for your previous question or not, this is the way I am using 

a=df.A.values[:,None]-df.A.values
np.fill_diagonal(a, 9999999)
np.argmin(abs(a),0)
Out[1160]: array([1, 0, 3, 2, 0], dtype=int64)

Upvotes: 1

Related Questions