Trennen einen Datenrahmens in mehrere Datenrahmen basierend auf dem Indexwert in Pandas

Question

Ich habe einen df wie unten dargestellt:

   DateTime  ModFlow(cfs)  ObsFlow(cfs) ModVol(f3) ObsVol(f3) 
Event 
Event 1  8/15/2016 15:35 11.85926   0   0.039530867 0 
Event 1  8/15/2016 10:05 30.05923   0   0.100197433 0 
Event 1  8/15/2016 10:00 31.10118   0   0.1036706  0 
Event 1  8/15/2016 9:55 32.17444   0   0.107248133 0 
Event 1  8/15/2016 4:10 0.6783166  0.5650155  0.002261055 0.001883385 
Event 10 6/23/2016 4:35 0.5573569  0.4814242  0.001857856 0.001604747 
Event 10 6/23/2016 4:40 0.5536903  0.3544892  0.001845634 0.001181631 
Event 10 6/23/2016 4:45 0.5502114  0.368421  0.001834038 0.00122807 
Event 10 6/23/2016 4:50 0.5698021  0.501548  0.00189934 0.001671827 
Event 10 6/23/2016 4:55 0.7525368  0.879257  0.002508456 0.002930857 
Event 11 6/10/2016 8:05 0.6593155  0.6145511  0.002197718 0.002048504 
Event 11 6/10/2016 8:10 0.6621117  0.8405573  0.002207039 0.002801858 
Event 11 6/10/2016 8:15 0.6578091  0.8173375  0.002192697 0.002724458 
Event 11 6/10/2016 8:20 0.6581948  0.871517  0.002193983 0.002905057 
Event 12 4/26/2016 22:00 2.307288  2.588235  0.00769096 0.00862745 
Event 12 4/26/2016 22:05 2.366998  3.091331  0.007889993 0.010304437 
Event 12 4/26/2016 22:10 2.494073  3.278638  0.008313577 0.010928793 
Event 12 4/26/2016 22:15 2.746868  3.083591  0.009156227 0.010278637 
Event 12 4/26/2016 22:20 3.146326  2.877709  0.010487753 0.009592363 
Event 12 4/26/2016 22:30 4.090476  2.354489  0.01363492 0.007848297 

ähnlich bis Ereignis 47

Ich mag für jedes Ereignis einen separaten Datenrahmen haben. Erwartete Ausgabe würde wie folgt aussehen:

Event1_df:

  DateTime  ModFlow(cfs) ObsFlow(cfs) ModVol(f3) Obs Vol(f3) 
Event     
     8/15/2016 15:35 11.85926  0    0.039530867 0 
     8/15/2016 10:05 30.05923  0    0.100197433 0 
Event 1 8/15/2016 10:00 31.10118  0    0.1036706  0 
     8/15/2016 9:55  32.17444  0    0.107248133 0 
     8/15/2016 4:10  0.6783166 0.5650155  0.002261055 0.001883385 

Event2_df:

  DateTime  ModFlow(cfs) ObsFlow(cfs) ModVol(f3) Obs Vol(f3) 
Event     
     8/15/2016 15:35 11.85926  0    0.039530867 0 
     8/15/2016 10:05 30.05923  0    0.100197433 0 
Event 2 8/15/2016 10:00 31.10118  0    0.1036706  0 
     8/15/2016 9:55  32.17444  0    0.107248133 0 
     8/15/2016 4:10  0.6783166 0.5650155  0.002261055 0.001883385 

Ähnliches gilt für andere Veranstaltungen

Wie könnte ich es tun?

Answer 1

Da Event der Index ist, können Sie loc verwenden, um Inhalte zu extrahieren und in dict zu platzieren.

In [482]: {x: df.loc[x] for x in df.index.unique()} 
Out[482]: 
{'Event1':   DateTime ModFlow(cfs) ObsFlow(cfs) ModVol(f3) ObsVol(f3) 
Event 
Event1 8/15/2016  11.859260  0.000000 0.039531 0.000000 
Event1 8/15/2016  30.059230  0.000000 0.100197 0.000000 
Event1 8/15/2016  31.101180  0.000000 0.103671 0.000000 
Event1 8/15/2016  32.174440  0.000000 0.107248 0.000000 
Event1 8/15/2016  0.678317  0.565016 0.002261 0.001883, 
'Event10':   DateTime ModFlow(cfs) ObsFlow(cfs) ModVol(f3) ObsVol(f3) Event 
Event10 6/23/2016  0.557357  0.481424 0.001858 0.001605 
Event10 6/23/2016  0.553690  0.354489 0.001846 0.001182 
Event10 6/23/2016  0.550211  0.368421 0.001834 0.001228 
Event10 6/23/2016  0.569802  0.501548 0.001899 0.001672 
Event10 6/23/2016  0.752537  0.879257 0.002508 0.002931, 
'Event11':   DateTime ModFlow(cfs) ObsFlow(cfs) ModVol(f3) ObsVol(f3) Event 
Event11 6/10/2016  0.659315  0.614551 0.002198 0.002049 
Event11 6/10/2016  0.662112  0.840557 0.002207 0.002802 
Event11 6/10/2016  0.657809  0.817338 0.002193 0.002724 
Event11 6/10/2016  0.658195  0.871517 0.002194 0.002905, 
'Event12':   DateTime ModFlow(cfs) ObsFlow(cfs) ModVol(f3) ObsVol(f3) Event 
Event12 4/26/2016  2.307288  2.588235 0.007691 0.008627 
Event12 4/26/2016  2.366998  3.091331 0.007890 0.010304 
Event12 4/26/2016  2.494073  3.278638 0.008314 0.010929 
Event12 4/26/2016  2.746868  3.083591 0.009156 0.010279 
Event12 4/26/2016  3.146326  2.877709 0.010488 0.009592 
Event12 4/26/2016  4.090476  2.354489 0.013635 0.007848} 

Details:

In [489]: df.index.unique() 
Out[489]: array(['Event1', 'Event10', 'Event11', 'Event12'], dtype=object) 

In [487]: df_split = {x: df.loc[x] for x in df.index.unique()} 

In [488]: df_split['Event1'] 
Out[488]: 
     DateTime ModFlow(cfs) ObsFlow(cfs) ModVol(f3) ObsVol(f3) 
Event 
Event1 8/15/2016  11.859260  0.000000 0.039531 0.000000 
Event1 8/15/2016  30.059230  0.000000 0.100197 0.000000 
Event1 8/15/2016  31.101180  0.000000 0.103671 0.000000 
Event1 8/15/2016  32.174440  0.000000 0.107248 0.000000 
Event1 8/15/2016  0.678317  0.565016 0.002261 0.001883 

Answer 2

Sie werden sie nicht in diesen Namen ohne Hartcodierung Variablennamen der Lage sein, zu vergeben. Sie können jedoch ein Wörterbuch erstellen, das für jeden Datenrahmen einen Schlüssel speichert. In diesem Fall ein Wörterbuch Verständnis:

> event_dict = {event: df.loc[event] for event in df.index.unique()} 

> event_dict['Event1'] 
# returns 
      DateTime ModFlow(cfs) ObsFlow(cfs) ModVol(f3) ObsVol(f3) 
Event 
Event1 8/15/2016  11.859260  0.000000 0.039531 0.000000 
Event1 8/15/2016  30.059230  0.000000 0.100197 0.000000 
Event1 8/15/2016  31.101180  0.000000 0.103671 0.000000 
Event1 8/15/2016  32.174440  0.000000 0.107248 0.000000 
Event1 8/15/2016  0.678317  0.565016 0.002261 0.001883