def dailyTimeDistributionFeatures ( dailyCallDistribution_dictionary, missingValue = -999, lowSampleValue = -666, numberOfFeatures = 14, sampleSizeThreshold = 3 ):
featureSelection = {}
for date in dailyCallDistribution_dictionary:
date_timestruct = datetime.datetime.fromtimestamp(time.mktime(time.strptime(date, "%Y-%m-%d")))
timeSample = dailyCallDistribution_dictionary[ date ]
if len(timeSample) <= sampleSizeThreshold :
if len(timeSample) == 0 :
featureSelection[ date ] = [ date_timestruct.timetuple().tm_wday
, int(date_timestruct.strftime('%W'))
, date_timestruct.month ] + [missingValue] * (numberOfFeatures - 3)
else :
featureSelection[ date ] = [ date_timestruct.timetuple().tm_wday
, int(date_timestruct.strftime('%W'))
, date_timestruct.month ] + [lowSampleValue] * (numberOfFeatures - 3)
else :
featureSelection[ date ] = [ date_timestruct.timetuple().tm_wday
, int(date_timestruct.strftime('%W'))
, date_timestruct.month
, len( timeSample )
# counts how many late night activities.
, sum( Pandas.Series( timeSample ).apply( lambda x: (x>0) & (x <= 4) ).tolist() )
, Pandas.Series( timeSample ).mean()
, Pandas.Series( timeSample ).median()
, Pandas.Series( timeSample ).std()
, Pandas.Series( timeSample ).min()
, Pandas.Series( timeSample ).max()
, Pandas.Series( timeSample ).mad()
, Pandas.Series( timeSample ).quantile(0.75) - Pandas.Series( timeSample ).quantile(0.25)
, Pandas.Series( timeSample ).kurt()
, Pandas.Series( timeSample ).skew()
]
return Pandas.DataFrame( featureSelection, index = ['dayOfWeek', 'WeekOfYear', 'MonthOfYear',
'Number of Calls', 'Number of Late Night Activities',
'Average Time', 'Median of Time',
'Standard Deviation', 'Earliest Call',
'Latest Call', 'Mean Absolute Deviation',
'Interquartile Range', 'Kurtosis',
'Skewness'] ).T
データフレームを出力し、上記の関数内のデータフレームにもう1列追加しようとする上記のpython関数を書いたとき：
featureTime['Whether Staying Late'] = featureTime['Number of Late Night Activities'].apply( lambda x: x > 0 ).apply(lambda x: sum(x))
エラーが発生しました：
TypeError Traceback (most recent call last)
/home/aaa/Enthought/Canopy_64bit/System/lib/python2.7/site- packages/IPython/utils/py3compat.pyc in execfile(fname, *where)
181 else:
182 filename = fname
--> 183 __builtin__.execfile(filename, *where)
/home/aaa/pyRepo/feature_selection_v15.py in <module>()
352 featureTime.to_csv('time.csv')
353
--> 354 featureTime['Whether Staying Late'] = featureTime['Number of Late Night Activities'].apply( lambda x: x > 0 ).apply(lambda x: sum(x))
355
356
/home/aaa/Enthought/Canopy_64bit/User/lib/python2.7/site- packages/pandas/core/series.pyc in apply(self, func, convert_dtype, args, **kwds)
2445 values = lib.map_infer(values, lib.Timestamp)
2446
-> 2447 mapped = lib.map_infer(values, f, convert=convert_dtype)
2448 if isinstance(mapped[0], Series):
2449 from pandas.core.frame import DataFrame
/home/aaa/Enthought/Canopy_64bit/User/lib/python2.7/site-packages/pandas/lib.so in pandas.lib.map_infer (pandas/lib.c:41822)()
/home/aaa/pyRepo/feature_selection_v15.py in <lambda>(x)
352 featureTime.to_csv('time.csv')
353
--> 354 featureTime['Whether Staying Late'] = featureTime['Number of Late Night Activities'].apply( lambda x: x > 0 ).apply(lambda x: sum(x))
355
356
TypeError: 'numpy.bool_' object is not iterable
コンソールに話しかけて手動で追加すると、存在しません。
代わりに Python の組み込みデータ型と for ループを使用して問題を解決しました。私が興味を持っているのは、なぜ上記のようなバグが発生するのかということです...それがどこから来たのか知りたいです...