def tree_to_table(tree):
"""Convert a ROOT TTree to an astropy Table.
Parameters
----------
Returns
-------
table : astropy.table.Table
"""
from rootpy import asrootpy
tree = asrootpy(tree)
array = tree.to_array()
from rootpy.io import open
from rootpy.root2array import tree_to_recarray
print('Reading %s' % infile)
file = open(infile)
tree_name = 'TableAllMu_WithoutNan' # 'ParTree_Postselect'
print('Getting %s' % tree_name)
tree = file.get(tree_name, ignore_unsupported=True)
print('Converting tree to recarray')
array = tree_to_recarray(tree)
print('Converting recarray to atpy.Table')
table = recarray_to_table(array)
print('Removing empty columns:')
empty_columns = []
for col in table.columns:
#if table['col'].min() == table['col'].max()
if (table['col'] == 0).all():
empty_columns.append(col)
print(empty_columns)
table.remove_columns(empty_columns)
for name in array.dtype.names:
# FITS can't save these types.
data = array[name]
if data.dtype == np.dtype('uint64'):
data = data.copy().astype('int64')
table.add_column(name, data)
return table
def tree_to_table(tree, tree_name):
"""Convert a ROOT TTree to an astropy Table.
Parameters
----------
tree : ROOT.TTree
ROOT TTree
Returns
-------
table : `~astropy.table.Table`
ROOT tree data as an astropy table.
"""
from rootpy import asrootpy
from rootpy.io import open
from rootpy.root2array import tree_to_recarray
tree = asrootpy(tree)
array = tree.to_array()
file = open(infile)
tree_name = 'TableAllMu_WithoutNan' # 'ParTree_Postselect'
tree = file.get(tree_name, ignore_unsupported=True)
array = tree_to_recarray(tree)
table = recarray_to_table(array)
# Remove empty columns
empty_columns = []
for col in table.columns:
#if table['col'].min() == table['col'].max()
if (table['col'] == 0).all():
empty_columns.append(col)
print(empty_columns)
table.remove_columns(empty_columns)
for name in array.dtype.names:
# FITS can't save these types.
data = array[name]
if data.dtype == np.dtype('uint64'):
data = data.copy().astype('int64')
table.add_column(name, data)
return table
table = None
ROOT.TTreeCache.SetLearnEntries(1)
time_in_read = 0
time_in_append = 0
processed_files = 0
try:
for file_chunk in chunk_files(flat_files, args.chainsize):
t1 = time.time()
chunk_chain = ROOT.TChain(tree)
chunk_chain.SetCacheSize(10000000)
for file in file_chunk:
chunk_chain.Add(file)
recarray = tree_to_recarray(chunk_chain, None, False)
t2 = time.time()
time_in_read += t2 - t1
t1 = time.time()
if table is None:
log.info("Creating table")
table = hfile.createTable(args.where, h5name,
recarray,
"", filters=filters,
expectedrows=entries)
else:
table.append(recarray)
table.flush()
t2 = time.time()
time_in_append += t2 - t1
print "time without profiler overhead:"
t1 = time.time()
arr1 = tree_to_recarray_py(tree, branches=branches,
include_weight=True)
t2 = time.time()
print "%f seconds" % (t2 - t1)
print '=' * 40
print "Using compiled C extension..."
cProfile.run('arr2 = tree_to_recarray(tree, branches=branches,'
'include_weight=True)')
print "time without profiler overhead:"
t1 = time.time()
arr2 = tree_to_recarray(tree, branches=branches,
include_weight=True)
t2 = time.time()
print "%f seconds" % (t2 - t1)
print '=' * 40
print "Comparison of output:"
print
print "Python result:"
print arr1
print arr1['a_x']
print arr1['weight']
print
print "C result:"
print arr2
# define the model
class Event(TreeModel):
x = FloatCol()
y = FloatCol()
z = FloatCol()
i = IntCol()
tree = Tree("test", model=Event)
# fill the tree
for i in xrange(100000):
tree.x = gauss(.5, 1.)
tree.y = gauss(.3, 2.)
tree.z = gauss(13., 42.)
tree.i = i
tree.fill()
tree.write()
# convert tree into a numpy record array
from rootpy.root2array import tree_to_recarray, tree_to_ndarray
array = tree_to_recarray(tree)
print array
print array.x
print array.i
print tree_to_ndarray(tree)
f.close()
table = None
ROOT.TTreeCache.SetLearnEntries(1)
time_in_read = 0
time_in_append = 0
processed_files = 0
try:
for file_chunk in chunk_files(flat_files, args.chainsize):
t1 = time.time()
chunk_chain = ROOT.TChain(tree)
chunk_chain.SetCacheSize(10000000)
for file in file_chunk:
chunk_chain.Add(file)
recarray = tree_to_recarray(chunk_chain, None, False)
t2 = time.time()
time_in_read += t2 - t1
t1 = time.time()
if table is None:
log.info("Creating table")
table = hfile.createTable(args.where,
h5name,
recarray,
"",
filters=filters,
expectedrows=entries)
else:
table.append(recarray)
table.flush()
print "time without profiler overhead:"
t1 = time.time()
arr1 = tree_to_recarray_py(tree, branches=branches, include_weight=True)
t2 = time.time()
print "%f seconds" % (t2 - t1)
print '=' * 40
print "Using compiled C extension..."
cProfile.run('arr2 = tree_to_recarray(tree, branches=branches,'
'include_weight=True)')
print "time without profiler overhead:"
t1 = time.time()
arr2 = tree_to_recarray(tree, branches=branches, include_weight=True)
t2 = time.time()
print "%f seconds" % (t2 - t1)
print '=' * 40
print "Comparison of output:"
print
print "Python result:"
print arr1
print arr1['a_x']
print arr1['weight']
print
print "C result:"
print arr2