def test_ensure_int32():
values = np.arange(10, dtype=np.int32)
result = com._ensure_int32(values)
assert (result.dtype == np.int32)
values = np.arange(10, dtype=np.int64)
result = com._ensure_int32(values)
assert (result.dtype == np.int32)
def test_ensure_int32():
values = np.arange(10, dtype=np.int32)
result = com._ensure_int32(values)
assert(result.dtype == np.int32)
values = np.arange(10, dtype=np.int64)
result = com._ensure_int32(values)
assert(result.dtype == np.int32)
def generate_groups(data, group_index, ngroups, axis=0, factory=lambda x: x):
"""
Parameters
----------
data : BlockManager
Returns
-------
generator
"""
group_index = com._ensure_int32(group_index)
indexer = lib.groupsort_indexer(group_index, ngroups)[0]
group_index = group_index.take(indexer)
if isinstance(data, BlockManager):
# this is sort of wasteful but...
sorted_axis = data.axes[axis].take(indexer)
sorted_data = data.reindex_axis(sorted_axis, axis=axis)
if isinstance(data, Series):
sorted_axis = data.index.take(indexer)
sorted_data = data.reindex(sorted_axis)
elif isinstance(data, DataFrame):
sorted_data = data.take(indexer, axis=axis)
if isinstance(sorted_data, DataFrame):
def _get_slice(slob):
if axis == 0:
return sorted_data[slob]
else:
return sorted_data.ix[:, slob]
elif isinstance(sorted_data, BlockManager):
def _get_slice(slob):
return factory(sorted_data.get_slice(slob, axis=axis))
elif isinstance(sorted_data, Series):
def _get_slice(slob):
return sorted_data._get_values(slob)
else: # pragma: no cover
def _get_slice(slob):
return sorted_data[slob]
starts, ends = lib.generate_slices(group_index, ngroups)
for i, (start, end) in enumerate(zip(starts, ends)):
# Since I'm now compressing the group ids, it's now not "possible" to
# produce empty slices because such groups would not be observed in the
# data
assert (start < end)
yield i, _get_slice(slice(start, end))
def group_info(self):
if len(self.groupings) > 1:
all_labels = [ping.labels for ping in self.groupings]
group_index = get_group_index(all_labels, self.shape)
comp_ids, obs_group_ids = _compress_group_index(group_index)
else:
ping = self.groupings[0]
group_index = ping.labels
comp_ids, obs_group_ids = _compress_group_index(group_index)
ngroups = len(obs_group_ids)
comp_ids = com._ensure_int32(comp_ids)
return comp_ids, obs_group_ids, ngroups
def group_info(self):
if len(self.groupings) > 1:
all_labels = [ping.labels for ping in self.groupings]
group_index = get_group_index(all_labels, self.shape)
comp_ids, obs_group_ids = _compress_group_index(group_index)
else:
ping = self.groupings[0]
group_index = ping.labels
comp_ids, obs_group_ids = _compress_group_index(group_index)
ngroups = len(obs_group_ids)
comp_ids = com._ensure_int32(comp_ids)
return comp_ids, obs_group_ids, ngroups
def generate_groups(data, group_index, ngroups, axis=0, factory=lambda x: x):
"""
Parameters
----------
data : BlockManager
Returns
-------
generator
"""
group_index = com._ensure_int32(group_index)
indexer = lib.groupsort_indexer(group_index, ngroups)[0]
group_index = group_index.take(indexer)
if isinstance(data, BlockManager):
# this is sort of wasteful but...
sorted_axis = data.axes[axis].take(indexer)
sorted_data = data.reindex_axis(sorted_axis, axis=axis)
if isinstance(data, Series):
sorted_axis = data.index.take(indexer)
sorted_data = data.reindex(sorted_axis)
elif isinstance(data, DataFrame):
sorted_data = data.take(indexer, axis=axis)
if isinstance(sorted_data, DataFrame):
def _get_slice(slob):
if axis == 0:
return sorted_data[slob]
else:
return sorted_data.ix[:, slob]
elif isinstance(sorted_data, BlockManager):
def _get_slice(slob):
return factory(sorted_data.get_slice(slob, axis=axis))
elif isinstance(sorted_data, Series):
def _get_slice(slob):
return sorted_data._get_values(slob)
else: # pragma: no cover
def _get_slice(slob):
return sorted_data[slob]
starts, ends = lib.generate_slices(group_index, ngroups)
for i, (start, end) in enumerate(zip(starts, ends)):
# Since I'm now compressing the group ids, it's now not "possible" to
# produce empty slices because such groups would not be observed in the
# data
assert(start < end)
yield i, _get_slice(slice(start, end))
def group_info(self):
comp_ids, obs_group_ids = self._get_compressed_labels()
ngroups = len(obs_group_ids)
comp_ids = com._ensure_int32(comp_ids)
return comp_ids, obs_group_ids, ngroups
