2014-12-22 35 views
8

Spróbuj tego dla siebie:Dlaczego DataFrame.loc [[1]] 1 800 x wolniej niż df.ix [[1]] i 3500x niż df.loc [1]?

import pandas as pd 
s=pd.Series(xrange(5000000)) 
%timeit s.loc[[0]] # You need pandas 0.15.1 or newer for it to be that slow 
1 loops, best of 3: 445 ms per loop 

Aktualizacja: który jest a legitimate bug in pandas że prawdopodobnie został wprowadzony w 0.15.1 w sierpniu 2014 roku lub tak. Obejścia: poczekaj na nowe wydanie, korzystając ze starej wersji pandy; zdobyć nowatorskiego dewelopera. wersja z github; ręcznie wykonaj jednoliniową modyfikację w wydaniu pandas; tymczasowo użyj .ix zamiast .loc.

mam DataFrame 4,8 miliona rzędach, i wybranie jednego rzędu przez .iloc[[ id ]] (z listy jednoelementowych) trwa 489 ms, prawie pół sekundy 1,800x razy wolniej niż identyczne .ix[[ id ]] i 3.500x razy wolniej niż.iloc[id] (przekazując identyfikator jako wartość, a nie jako listę). Aby być uczciwym, .loc[list] zajmuje mniej więcej ten sam czas, niezależnie od długości listy, ale nie chcę wydać na nim 489 ms, zwłaszcza gdy .ix jest tysiąc razy szybszy i daje taki sam wynik. Rozumiem, że powinno być wolniej, nieprawdaż?

Używam pand 0.15.1. Doskonały samouczek na temat Indexing and Selecting Data sugeruje, że .ix jest jakoś bardziej ogólny i prawdopodobnie wolniejszy niż .loc i .iloc. Konkretnie, to mówi

Jednak, gdy oś jest liczbą całkowitą oparte tylko dostęp w oparciu etykiety i nie pozycyjną dostępu jest obsługiwana. Dlatego w takich przypadkach zwykle lepiej jest jawnie i użyć .iloc lub .loc.

Oto sesja ipython z benchmarków:

print 'The dataframe has %d entries, indexed by integers that are less than %d' % (len(df), max(df.index)+1) 
    print 'df.index begins with ', df.index[:20] 
    print 'The index is sorted:', df.index.tolist()==sorted(df.index.tolist()) 

    # First extract one element directly. Expected result, no issues here. 
    id=5965356 
    print 'Extract one element with id %d' % id 
    %timeit df.loc[id] 
    %timeit df.ix[id] 
    print hash(str(df.loc[id])) == hash(str(df.ix[id])) # check we get the same result 

    # Now extract this one element as a list. 
    %timeit df.loc[[id]] # SO SLOW. 489 ms vs 270 microseconds for .ix, or 139 microseconds for .loc[id] 
    %timeit df.ix[[id]] 
    print hash(str(df.loc[[id]])) == hash(str(df.ix[[id]])) # this one should be True 
    # Let's double-check that in this case .ix is the same as .loc, not .iloc, 
    # as this would explain the difference. 
    try: 
     print hash(str(df.iloc[[id]])) == hash(str(df.ix[[id]])) 
    except: 
     print 'Indeed, %d is not even a valid iloc[] value, as there are only %d rows' % (id, len(df)) 

    # Finally, for the sake of completeness, let's take a look at iloc 
    %timeit df.iloc[3456789] # this is still 100+ times faster than the next version 
    %timeit df.iloc[[3456789]] 

wyjściowa:

The dataframe has 4826616 entries, indexed by integers that are less than 6177817 
df.index begins with Int64Index([1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20], dtype='int64') 
The index is sorted: True 
Extract one element with id 5965356 
10000 loops, best of 3: 139 µs per loop 
10000 loops, best of 3: 141 µs per loop 
True 
1 loops, best of 3: 489 ms per loop 
1000 loops, best of 3: 270 µs per loop 
True 
Indeed, 5965356 is not even a valid iloc[] value, as there are only 4826616 rows 
10000 loops, best of 3: 98.9 µs per loop 
100 loops, best of 3: 12 ms per loop 
+0

Należy zauważyć, że za pomocą '[[id]] 'i' [id]' nie są równoważne. '[id]' zwróci Serię, ale '[[id]]' zwróci jednorzędowe DataFrame. – BrenBarn

+0

@BrenBarn, tak, to wyjaśnia różnicę dla '.ix': 141 μs vs. 270 μs. Ale dlaczego ".loc [[id]]" jest tak wolny? – osa

Odpowiedz

6

Wygląda na to, że kwestia ta nie była obecna w pand 0.14. Profilowałem go pod numerem line_profiler i myślę, że wiem, co się stało. Od pandy 0.15.1, teraz KeyError jest podnoszona, jeśli dany indeks nie jest obecny. Wygląda na to, że gdy używasz składni .loc[list], wykonuje on wyczerpujące wyszukiwanie indeksu wzdłuż całej osi, nawet jeśli zostało znalezione. Oznacza to, po pierwsze, że nie ma wcześniejszego rozwiązania w przypadku znalezienia elementu, a po drugie, wyszukiwanie w tym przypadku jest brutalne życie.

File: .../anaconda/lib/python2.7/site-packages/pandas/core/indexing.py,

1278              # require at least 1 element in the index 
    1279   1   241 241.0  0.1    idx = _ensure_index(key) 
    1280   1  391040 391040.0  99.9    if len(idx) and not idx.isin(ax).any(): 
    1281           
    1282               raise KeyError("None of [%s] are in the [%s]" % 
4

Pandy indeksowania szalony wolno, że przełącza się NumPy indeksowania

df=pd.DataFrame(some_content) 
# takes forever!! 
for iPer in np.arange(-df.shape[0],0,1): 
    x = df.iloc[iPer,:].values 
    y = df.iloc[-1,:].values 
# fast!   
vals = np.matrix(df.values) 
for iPer in np.arange(-vals.shape[0],0,1): 
    x = vals[iPer,:] 
    y = vals[-1,:]