Alternatywa pandas dla pętli

Wydajny kod w Pythonie

Logan Thomas

Scientific Software Technical Trainer, Enthought

print(baseball_df.head())

  Team League  Year   RS   RA   W    G  Playoffs
0  ARI     NL  2012  734  688  81  162         0
1  ATL     NL  2012  700  600  94  162         1
2  BAL     AL  2012  712  705  93  162         1
3  BOS     AL  2012  734  806  69  162         0
4  CHC     NL  2012  613  759  61  162         0

def calc_run_diff(runs_scored, runs_allowed):

    run_diff = runs_scored - runs_allowed

    return run_diff

Różnice przebiegów z pętlą

run_diffs_iterrows = []

for i,row in baseball_df.iterrows():
    run_diff = calc_run_diff(row['RS'], row['RA'])
    run_diffs_iterrows.append(run_diff)

baseball_df['RD'] = run_diffs_iterrows
print(baseball_df)

     Team League  Year   RS   RA    W    G  Playoffs   RD
0     ARI     NL  2012  734  688   81  162         0   46
1     ATL     NL  2012  700  600   94  162         1  100
2     BAL     AL  2012  712  705   93  162         1    7
...

Metoda .apply() w pandas

Przyjmuje funkcję i stosuje ją do DataFrame
- Należy określić oś (0 dla kolumn; 1 dla wierszy)
Można używać z funkcjami anonimowymi (lambda)
Przykład:

baseball_df.apply(

    lambda row: calc_run_diff(row['RS'], row['RA']),

    axis=1
)

Różnice przebiegów z .apply()

run_diffs_apply = baseball_df.apply(
         lambda row: calc_run_diff(row['RS'], row['RA']),
         axis=1)

baseball_df['RD'] = run_diffs_apply
print(baseball_df)

     Team League  Year   RS   RA    W    G  Playoffs   RD
0     ARI     NL  2012  734  688   81  162         0   46
1     ATL     NL  2012  700  600   94  162         1  100
2     BAL     AL  2012  712  705   93  162         1    7
...

Porównanie podejść

%%timeit
run_diffs_iterrows = []

for i,row in baseball_df.iterrows():
    run_diff = calc_run_diff(row['RS'], row['RA'])
    run_diffs_iterrows.append(run_diff)

baseball_df['RD'] = run_diffs_iterrows

86.8 ms ± 3 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)

Porównanie podejść

%%timeit
run_diffs_apply = baseball_df.apply(
         lambda row: calc_run_diff(row['RS'], row['RA']),
         axis=1)

baseball_df['RD'] = run_diffs_apply

30.1 ms ± 1.75 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)

Ćwiczmy metodę .apply() w pandas!

Wydajny kod w Pythonie