def create_corpus(self, table_name):
self.odps.execute_sql("drop table if exists " + table_name)
self.odps.execute_sql("create table %s (id string, content string) lifecycle 3" % table_name)
upload_ss = self.tunnel.create_upload_session(table_name)
writer = upload_ss.open_record_writer(0)
for line_no, line in enumerate(load_resource_string('odps.examples.data', 'splited_words.txt').splitlines()):
rec = upload_ss.new_record()
cols = [line_no + 1, line.replace('####', '')]
[rec.set(i, val) for i, val in enumerate(cols)]
writer.write(rec)
writer.close()
upload_ss.commit([0, ])
def create_iris_kv(self, table_name):
self.odps.execute_sql("drop table if exists " + table_name)
self.odps.execute_sql('create table %s (content string, category bigint) lifecycle 3' % table_name)
upload_ss = self.tunnel.create_upload_session(table_name)
writer = upload_ss.open_record_writer(0)
for line in load_resource_string('odps.examples.data', 'iris.txt').splitlines():
rec = upload_ss.new_record()
line_parts = line.split(',')
rec.set(0, ','.join('%s:%s' % (idx, c) for idx, c in enumerate(line_parts[:-1])))
rec.set(1, 0 if 'setosa' in line_parts[-1] else 1)
writer.write(rec)
writer.close()
upload_ss.commit([0, ])
def create_ionosphere(self, table_name):
fields = ','.join('a%02d double' % i for i in range(1, 35)) + ', class bigint'
self.odps.execute_sql("drop table if exists " + table_name)
self.odps.execute_sql("create table %s (%s) lifecycle 3" % (table_name, fields))
upload_ss = self.tunnel.create_upload_session(table_name)
writer = upload_ss.open_record_writer(0)
for line in load_resource_string('odps.examples.data', 'ionosphere.txt').splitlines():
rec = upload_ss.new_record()
cols = [float(c) if rec._columns[i].type == 'double' else int(c) for i, c in enumerate(line.split(','))]
[rec.set(i, val) for i, val in enumerate(cols)]
writer.write(rec)
writer.close()
upload_ss.commit([0, ])
def create_iris(self, table_name):
self.odps.execute_sql("drop table if exists " + table_name)
self.odps.execute_sql(('create table %s (sepal_length double, sepal_width double, petal_length double, '
+ 'petal_width double, category string) lifecycle 3') % table_name)
upload_ss = self.tunnel.create_upload_session(table_name)
writer = upload_ss.open_record_writer(0)
for line in load_resource_string('odps.examples.data', 'iris.txt').splitlines():
rec = upload_ss.new_record()
line_parts = line.split(',')
cols = [float(c) for c in line_parts[:-1]]
cols.append(line_parts[4])
[rec.set(i, val) for i, val in enumerate(cols)]
writer.write(rec)
writer.close()
upload_ss.commit([0, ])
def create_splited_words(self, table_name):
self.odps.execute_sql("drop table if exists " + table_name)
self.odps.execute_sql("create table %s (id string, content string) lifecycle 3" % table_name)
upload_ss = self.tunnel.create_upload_session(table_name)
writer = upload_ss.open_record_writer(0)
for line_no, line in enumerate(load_resource_string('odps.examples.data', 'splited_words.txt').splitlines()):
if not line.strip():
break
for word in line.split('####'):
rec = upload_ss.new_record()
cols = [line_no + 1, word]
[rec.set(i, val) for i, val in enumerate(cols)]
writer.write(rec)
writer.close()
upload_ss.commit([0, ])
def create_word_triple(self, table_name):
self.odps.execute_sql("drop table if exists " + table_name)
self.odps.execute_sql("create table %s (id string, word string, count bigint) lifecycle 3" % table_name)
upload_ss = self.tunnel.create_upload_session(table_name)
writer = upload_ss.open_record_writer(0)
for line_no, line in enumerate(load_resource_string('odps.examples.data', 'splited_words.txt').splitlines()):
line = line.strip()
if not line:
break
for word, group in groupby(sorted(line.split('####'))):
rec = upload_ss.new_record()
cols = [str(line_no + 1), word, len(list(group))]
[rec.set(i, val) for i, val in enumerate(cols)]
writer.write(rec)
writer.close()
upload_ss.commit([0, ])
def create_ionosphere_one_part(self, table_name, partition_count=3):
fields = ','.join('a%02d double' % i for i in range(1, 35)) + ', class bigint'
self.odps.execute_sql("drop table if exists " + table_name)
self.odps.execute_sql("create table %s (%s) partitioned by (part bigint) lifecycle 3" % (table_name, fields))
for part_id in range(partition_count):
self.odps.execute_sql('alter table %s add if not exists partition (part=%d)' % (table_name, part_id))
upload_sses = [self.tunnel.create_upload_session(table_name, 'part=%d' % part_id) for part_id in range(partition_count)]
writers = [session.open_record_writer(0) for session in upload_sses]
for line_no, line in enumerate(load_resource_string('odps.examples.data', 'ionosphere.txt').splitlines()):
part_id = line_no % partition_count
rec = upload_sses[part_id].new_record()
cols = [float(c) if rec._columns[i].type == 'double' else int(c) for i, c in enumerate(line.split(','))]
cols.append(part_id)
[rec.set(i, val) for i, val in enumerate(cols)]
writers[part_id].write(rec)
[writer.close() for writer in writers]
[upload_ss.commit([0, ]) for upload_ss in upload_sses]
def create_ionosphere_two_parts(self, table_name, partition1_count=2, partition2_count=3):
fields = ','.join('a%02d double' % i for i in range(1, 35)) + ', class bigint'
self.odps.execute_sql("drop table if exists " + table_name)
self.odps.execute_sql("create table %s (%s) partitioned by (part1 bigint, part2 bigint) lifecycle 3" % (table_name, fields))
for id1, id2 in product(range(partition1_count), range(partition2_count)):
self.odps.execute_sql('alter table %s add if not exists partition (part1=%d, part2=%d)' % (table_name, id1, id2))
upload_sses = [[self.tunnel.create_upload_session(table_name, 'part1=%d,part2=%d' % (id1, id2))
for id2 in range(partition2_count)] for id1 in range(partition1_count)]
writers = [[session.open_record_writer(0) for session in sessions] for sessions in upload_sses]
for line_no, line in enumerate(load_resource_string('odps.examples.data', 'ionosphere.txt').splitlines()):
id1, id2 = line_no % partition1_count, line_no % partition2_count
rec = upload_sses[id1][id2].new_record()
cols = [float(c) if rec._columns[i].type == 'double' else int(c) for i, c in enumerate(line.split(','))]
cols.extend([id1, id2])
[rec.set(i, val) for i, val in enumerate(cols)]
writers[id1][id2].write(rec)
[writer.close() for ws in writers for writer in ws]
[upload_ss.commit([0, ]) for upload_sss in upload_sses for upload_ss in upload_sss]
