合并(Merge)

The Concat docs. The Join docs.

Append two dataframes with overlapping index (emulate R rbind)

  1. In [149]: rng = pd.date_range('2000-01-01', periods=6)
  3. In [150]: df1 = pd.DataFrame(np.random.randn(6, 3), index=rng, columns=['A', 'B', 'C'])
  5. In [151]: df2 = df1.copy()

Depending on df construction, ignore_index may be needed

  1. In [152]: df = df1.append(df2,ignore_index=True); df
  2. Out[152]: 
  3.            A         B         C
  4. 0  -0.480676 -1.305282 -0.212846
  5. 1   1.979901  0.363112 -0.275732
  6. 2  -1.433852  0.580237 -0.013672
  7. 3   1.776623 -0.803467  0.521517
  8. 4  -0.302508 -0.442948 -0.395768
  9. 5  -0.249024 -0.031510  2.413751
  10. 6  -0.480676 -1.305282 -0.212846
  11. 7   1.979901  0.363112 -0.275732
  12. 8  -1.433852  0.580237 -0.013672
  13. 9   1.776623 -0.803467  0.521517
  14. 10 -0.302508 -0.442948 -0.395768
  15. 11 -0.249024 -0.031510  2.413751

Self Join of a DataFrame

  1. In [153]: df = pd.DataFrame(data={'Area' : ['A'] * 5 + ['C'] * 2,
  2.    .....:                         'Bins' : [110] * 2 + [160] * 3 + [40] * 2,
  3.    .....:                         'Test_0' : [0, 1, 0, 1, 2, 0, 1],
  4.    .....:                         'Data' : np.random.randn(7)});df
  5.    .....: 
  6. Out[153]: 
  7.   Area  Bins  Test_0      Data
  8. 0    A   110       0 -0.378914
  9. 1    A   110       1 -1.032527
  10. 2    A   160       0 -1.402816
  11. 3    A   160       1  0.715333
  12. 4    A   160       2 -0.091438
  13. 5    C    40       0  1.608418
  14. 6    C    40       1  0.753207
  16. In [154]: df['Test_1'] = df['Test_0'] - 1
  18. In [155]: pd.merge(df, df, left_on=['Bins', 'Area','Test_0'], right_on=['Bins', 'Area','Test_1'],suffixes=('_L','_R'))
  19. Out[155]: 
  20.   Area  Bins  Test_0_L    Data_L  Test_1_L  Test_0_R    Data_R  Test_1_R
  21. 0    A   110         0 -0.378914        -1         1 -1.032527         0
  22. 1    A   160         0 -1.402816        -1         1  0.715333         0
  23. 2    A   160         1  0.715333         0         2 -0.091438         1
  24. 3    C    40         0  1.608418        -1         1  0.753207         0

How to set the index and join

KDB like asof join

Join with a criteria based on the values

Using searchsorted to merge based on values inside a range