Skip to content

Instantly share code, notes, and snippets.

@mesmacosta
Last active January 6, 2020 14:04
Show Gist options
  • Select an option

  • Save mesmacosta/86bd6a4b4938eaf67dd2d95a3afa3ef9 to your computer and use it in GitHub Desktop.

Select an option

Save mesmacosta/86bd6a4b4938eaf67dd2d95a3afa3ef9 to your computer and use it in GitHub Desktop.
import argparse
import logging
import random
import sys
import uuid
from teradatasql import connect
_DATA_TYPES = [
'INTEGER', 'CHAR(25)', 'DECIMAL(8,2)', 'DATE', 'VARCHAR(25)', 'SMALLINT',
'CHAR', 'BYTEINT'
]
_COLUMN_NAMES = [
'name', 'address', 'city', 'state', 'date_time', 'paragraph', 'randomdata',
'person', 'credit_card', 'size', 'reason', 'school', 'food', 'location',
'house', 'price', 'cpf', 'cnpj', 'passport', 'security_number',
'phone_number', 'bank_account_number', 'ip_address', 'stocks'
]
_TABLE_NAMES = [
'school_info', 'personal_info', 'persons', 'employees', 'companies',
'store', 'home'
]
_DATABASE_NAMES = [
'school_warehouse', 'company_warehouse', 'on_prem_warehouse',
'factory_warehouse', 'organization_warehouse'
]
def get_conn(connection_args):
return connect(None,
database=connection_args['database'],
host=connection_args['host'],
user=connection_args['user'],
password=connection_args['pass'])
def create_random_metadata(connection_args):
conn = get_conn(connection_args)
cursor = conn.cursor()
for x in range(4):
database_name, database_stmt = build_create_database_statement()
cursor.execute(database_stmt)
for y in range(250):
query = build_create_table_statement(database_name)
print('\n' + query)
cursor.execute(query)
conn.commit()
cursor.close()
def get_random_database_name():
return random.choice(_DATABASE_NAMES)
def build_create_database_statement():
database_name = '{}{}'.format(get_random_database_name(),
str(random.randint(1, 100000)))
database_stmt = 'CREATE DATABASE "{}" AS '.format(
database_name)
database_stmt += 'PERM = 2000000*(HASHAMP()+1), '
database_stmt += 'SPOOL = 2000000*(HASHAMP()+1), '
database_stmt += 'TEMPORARY = 2000000*(HASHAMP()+1) '
return database_name, database_stmt
def get_random_data_type():
return random.choice(_DATA_TYPES)
def get_random_column_name():
return random.choice(_COLUMN_NAMES)
def get_random_table_name():
return random.choice(_TABLE_NAMES)
def build_create_table_statement(database_name):
table_stmt = 'CREATE TABLE {}.{}{} ( '.format(database_name,
get_random_table_name(),
uuid.uuid4().hex[:8])
table_stmt = '{}{}{} {}'.format(table_stmt, get_random_column_name(),
str(random.randint(1, 100000)),
get_random_data_type())
for x in range(random.randint(1, 15)):
table_stmt += ', {}{}'.format(get_random_column_name(),
str(random.randint(1, 100000))) + \
' {}'.format(get_random_data_type())
table_stmt = '{} )'.format(table_stmt)
return table_stmt
def parse_args():
parser = argparse.ArgumentParser(
description='Command line generate random metadata into teradata')
parser.add_argument('--teradata-host',
help='Your teradata server host',
required=True)
parser.add_argument('--teradata-user',
help='Your teradata credentials user')
parser.add_argument('--teradata-pass',
help='Your teradata credentials password')
parser.add_argument('--teradata-database',
help='Your teradata database name')
return parser.parse_args()
if __name__ == "__main__":
args = parse_args()
# Enable logging
logging.basicConfig(stream=sys.stdout, level=logging.DEBUG)
create_random_metadata({
'database': args.teradata_database,
'host': args.teradata_host,
'user': args.teradata_user,
'pass': args.teradata_pass
})
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment