def __getitem__(self, val):
"""Returns a TimeSeriesRDD representing a subslice of this TimeSeriesRDD, containing only
values for a sub-range of the time it covers.
"""
start = datetime_to_millis(val.start)
stop = datetime_to_millis(val.stop)
return TimeSeriesRDD(None, None, self._jtsrdd.slice(start, stop), self.ctx)
def uniform(start, end=None, periods=None, freq=None, sc=None):
"""
Instantiates a uniform DateTimeIndex.
Either end or periods must be specified.
Parameters
----------
start : string, long (millis from epoch), or Pandas Timestamp
end : string, long (millis from epoch), or Pandas Timestamp
periods : int
freq : a frequency object
sc : SparkContext
"""
dtmodule = sc._jvm.com.cloudera.sparkts.__getattr__(
'DateTimeIndex$').__getattr__('MODULE$')
if freq is None:
raise ValueError("Missing frequency")
elif end is None and periods == None:
raise ValueError("Need an end date or number of periods")
elif end is not None:
return DateTimeIndex(dtmodule.uniform( \
datetime_to_millis(start), datetime_to_millis(end), freq._jfreq))
else:
return DateTimeIndex(dtmodule.uniform( \
datetime_to_millis(start), periods, freq._jfreq))
def uniform(start, end=None, periods=None, freq=None, sc=None):
"""
Instantiates a uniform DateTimeIndex.
Either end or periods must be specified.
Parameters
----------
start : string, long (millis from epoch), or Pandas Timestamp
end : string, long (millis from epoch), or Pandas Timestamp
periods : int
freq : a frequency object
sc : SparkContext
"""
dtmodule = sc._jvm.com.cloudera.sparkts.__getattr__('DateTimeIndex$').__getattr__('MODULE$')
if freq is None:
raise ValueError("Missing frequency")
elif end is None and periods == None:
raise ValueError("Need an end date or number of periods")
elif end is not None:
return DateTimeIndex(dtmodule.uniform( \
datetime_to_millis(start), datetime_to_millis(end), freq._jfreq))
else:
return DateTimeIndex(dtmodule.uniform( \
datetime_to_millis(start), periods, freq._jfreq))
def __getitem__(self, val):
# TODO: throw an error if the step size is defined
if isinstance(val, slice):
start = datetime_to_millis(val.start)
stop = datetime_to_millis(val.stop)
jdt_index = self._jdt_index.slice(start, stop)
return DateTimeIndex(jdt_index)
else:
return self._jdt_index.locAtDateTime(datetime_to_millis(val))
def __getitem__(self, val):
"""
Returns a TimeSeriesRDD representing a subslice of this TimeSeriesRDD, containing only
values for a sub-range of the time it covers.
"""
start = datetime_to_millis(val.start)
stop = datetime_to_millis(val.stop)
return TimeSeriesRDD(None, None, self._jtsrdd.slice(start, stop),
self.ctx)
def __getitem__(self, val):
# TODO: throw an error if the step size is defined
if isinstance(val, slice):
start = datetime_to_millis(val.start)
stop = datetime_to_millis(val.stop)
jdt_index = self._jdt_index.slice(start, stop)
return DateTimeIndex(jdt_index)
else:
return self._jdt_index.locAtDateTime(datetime_to_millis(val))
def irregular(timestamps, sc):
"""
Instantiates an irregular DateTimeIndex.
Parameters
----------
timestamps : a Pandas DateTimeIndex, or an array of strings, longs (nanos from epoch), Pandas
Timestamps
sc : SparkContext
"""
dtmodule = sc._jvm.com.cloudera.sparkts.__getattr__('DateTimeIndex$').__getattr__('MODULE$')
arr = sc._gateway.new_array(sc._jvm.long, len(timestamps))
for i in xrange(len(timestamps)):
arr[i] = datetime_to_millis(timestamps[i])
return DateTimeIndex(dtmodule.irregular(arr))
def irregular(timestamps, sc):
"""
Instantiates an irregular DateTimeIndex.
Parameters
----------
timestamps : a Pandas DateTimeIndex, or an array of strings, longs (nanos from epoch), Pandas
Timestamps
sc : SparkContext
"""
dtmodule = sc._jvm.com.cloudera.sparkts.__getattr__(
'DateTimeIndex$').__getattr__('MODULE$')
arr = sc._gateway.new_array(sc._jvm.long, len(timestamps))
for i in xrange(len(timestamps)):
arr[i] = datetime_to_millis(timestamps[i])
return DateTimeIndex(dtmodule.irregular(arr))