Skip to content

B
bcom-tp-etl-transformation-pipelines
Commit b4835849 authored by Cristian Aguirre's avatar Cristian Aguirre
Browse files
Options

Merge remote-tracking branch 'origin/developer_ev' into developer_ca 

# Conflicts:
#	dags/procedure_prueba.json
parents c4edffad 852ec9f6
Show whitespace changes
Inline Side-by-side
Showing with 112 additions and 64 deletions
dags/components/Cleaning.py
View file @ b4835849
......@@ -18,13 +18,16 @@ import logging
logger = logging.getLogger()
def validate_clean(control_params: Dict[str, Any], provider: str, timezone: str, engine, **kwargs) -> None:
def validate_clean(control_params: Dict[str, Any], provider: str, timezone: str, intern_conn, **kwargs) -> None:
engine = intern_conn.engine
delete_task_instances()
ti = kwargs["ti"]
#created_Tables = ti.xcom_pull(task_ids="VALIDATE_GENERATOR", key="CREATED_TABLES")
procedures = ti.xcom_pull(task_ids="MASTER_TRANSFORMATION", key="PROC_CREATED")
if procedures:
for procedure in procedures:
with engine.connect() as conn:
result_p = bool(intern_conn.check_procedure(procedure, conn))
if result_p:
logger.info(f"Borrando procedures {procedure}")
delete = delete_procedure(procedure, engine)
if delete:
......@@ -60,12 +63,22 @@ def validate_clean(control_params: Dict[str, Any], provider: str, timezone: str,
def clean(command: str, intern_conn):
engine = intern_conn.engine
#tablename = select_multiple(command)["tablename"]
tablename = command
logger.info(f"Borrando tabla {tablename}")
delete = delete_table(tablename, engine)
if delete:
# with engine.connect() as conn:
# result_t = bool(intern_conn.check_table(tablename, conn))
# result_p = bool(intern_conn.check_procedure(tablename, conn))
# if result_t:
# logger.info(f"Borrando tabla {tablename}")
# delete_t = delete_table(tablename, engine)
# elif result_p:
# logger.info(f"Borrando procedure {tablename}")
# delete_p = delete_procedure(tablename, engine)
delete_p = delete_procedure(tablename, engine)
delete_t = delete_table(tablename, engine)
if delete_t:
logger.info(f"Borrado correctamente la tabla {tablename}")
elif delete_p:
logger.info(f"Borrado correctamente el procedure {tablename}")
logger.info(f"Borrado todas las variables xcom")
......@@ -110,7 +123,7 @@ def get_cleaning_task_group(db_intern_conn, control_s3: Dict[str, Any], provider
validate_task = PythonOperator(
task_id="VALIDATE_CLEANER",
python_callable=validate_clean,
op_kwargs={'control_params': control_s3, 'provider': provider, 'timezone': timezone, 'engine':db_intern_conn.engine},
op_kwargs={'control_params': control_s3, 'provider': provider, 'timezone': timezone, 'intern_conn':db_intern_conn},
trigger_rule='none_skipped'
)
cleaners >> tasks >> validate_task
......
dags/components/DatabaseOperation/DatabaseTransformation.py
View file @ b4835849
import logging
import time
from typing import List
import re
logger = logging.getLogger()
......@@ -16,17 +16,33 @@ def execute_transformations(commands: List[str], engine):
def delete_table(tablename: str, engine) -> bool:
base_Datos = engine.dialect.name
delete = False
try:
command = f'DROP TABLE IF EXISTS {tablename}'
if base_Datos == "oracle":
command = f'DROP TABLE {tablename}'
start_time = time.time()
with engine.connect() as conn:
try:
_ = conn.execute(command)
check_query = f"SELECT table_name FROM all_tables WHERE table_name = '{tablename.upper()}'"
result = conn.execute(check_query)
exists = bool(result.fetchone())
except Exception as e:
logger.error(f"Tabla no encontrada. {e}")
if exists:
_ = conn.execute(command)
delete = True
logger.debug(f"Duración de borrado: {time.time() - start_time}")
elif base_Datos == "mysql":
command = f'DROP TABLE IF EXISTS {tablename}'
start_time = time.time()
with engine.connect() as conn:
try:
_ = conn.execute(command)
delete = True
logger.debug(f"Duración de borrado: {time.time() - start_time}")
except Exception as e:
logger.error(f"Tabla no encontrada. {e}")
except Exception as e:
logger.error(f"Error borrando tabla {tablename}. {e}")
finally:
......@@ -34,17 +50,34 @@ def delete_table(tablename: str, engine) -> bool:
def delete_procedure(procedure: str, engine) -> bool:
base_Datos = engine.dialect.name
delete = False
try:
command = f"DROP PROCEDURE IF EXISTS {procedure}"
if base_Datos == "oracle":
command = f"DROP PROCEDURE {procedure}"
start_time = time.time()
with engine.connect() as conn:
try:
_ = conn.execute(command)
proc = re.match(r'([^(]+)', procedure).group(1)
check_query = f"SELECT object_name FROM all_procedures WHERE object_name = '{proc.upper()}'"
result = conn.execute(check_query)
exists = bool(result.fetchone())
except Exception as e:
logger.error(f"Procedure no encontrado. {e}")
if exists:
_ = conn.execute(command)
delete = True
logger.debug(f"Duración de borrado: {time.time() - start_time}")
elif base_Datos == "mysql":
command = f'DROP PROCEDURE IF EXISTS {procedure}'
start_time = time.time()
with engine.connect() as conn:
try:
_ = conn.execute(command)
delete = True
logger.debug(f"Duración de borrado: {time.time() - start_time}")
except Exception as e:
logger.error(f"Procedure no encontrada. {e}")
except Exception as e:
logger.error(f"Error borrando procedure {procedure}. {e}")
finally:
......
dags/components/Databases/Oracle.py
View file @ b4835849
......@@ -2,7 +2,7 @@ from typing import List, Tuple
from sqlalchemy import create_engine
import oracledb
import re
from enums.DatabaseDialectEnum import DatabaseDialectEnum
from enums.DatabaseTypeEnum import DatabaseTypeEnum
from components.Databases.Enums.OracleDataTypeEnum import OracleDataTypeEnum
......@@ -90,8 +90,9 @@ class Oracle:
def generate_sql_procedure(self, command: str, reserved_word: str = "execute") -> str:
response = ""
try:
logger.info("COMANDO",command)
command = command.replace(reserved_word, "").replace(";", "")
response = f"begin {command}; end;"
response = f"BEGIN {command}; END;"
logger.debug("COMANDO ORACLE:", response)
except Exception as e:
logger.error(f"Error generando comando sql para procedure Oracle. Comando: {command}. {e}")
......@@ -113,7 +114,8 @@ class Oracle:
def check_procedure(self, procedure_name, connection) -> bool:
exists = False
try:
check_query = f"SELECT text FROM all_source WHERE name = '{procedure_name}'"
procedure_name = re.match(r'([^(]+)', procedure_name).group(1)
check_query = f"SELECT object_name FROM all_procedures WHERE object_name = '{procedure_name.upper()}'"
result = connection.execute(check_query)
exists = bool(result.fetchone())
except Exception as e:
......@@ -124,7 +126,7 @@ class Oracle:
def check_table(self, table_name, connection) -> bool:
exists = False
try:
check_query = f"DESCRIBE {table_name}"
check_query = f"SELECT table_name FROM all_tables WHERE table_name = '{table_name.upper()}'"
result = connection.execute(check_query)
exists = bool(result.fetchone())
except Exception as e:
......@@ -135,7 +137,7 @@ class Oracle:
def verify_table(self, table_name, connection) -> bool:
exists = False
try:
check_query = f"SELECT COUNT(*) FROM {table_name}"
check_query = f"SELECT COUNT(*) FROM {table_name.upper()}"
result = connection.execute(check_query)
if result > 0:
exists = True
......
dags/dag_conf.yml
View file @ b4835849
......@@ -5,23 +5,23 @@ app:
database:
sources:
source1:
type: mysql
host: 192.168.1.13
port: 13306
username: root
password: root
database: prueba
service:
schema: sources
type: oracle
host: 192.168.27.22
port: 21521
username: PRUEBABCOM2
password: admin
database:
service: ORCLPDB1
schema:
transformation:
type: mysql
host: 192.168.1.13
port: 13306
username: root
password: root
database: prueba_ca
service:
schema: intern_db
type: oracle
host: 192.168.27.22
port: 21521
username: RLQA_AIR
password: RLQA_AIR99
database:
service: ORCLPDB1
schema:
chunksize: 4000
label_multiple_select: TABLENAME
label_transform_procedure: STORE
......@@ -55,14 +55,14 @@ app:
delimiter: '|'
tmp_path: /tmp
s3_params:
tabla4:
bucket: pruebairflow
tabla1:
bucket: prueba-id
prefix: bcom_results
connection_id: prueba_af
connection_id: conn_script
tabla5:
bucket: pruebairflow
bucket: prueba-id
prefix: bcom_results
connection_id: prueba_af
connection_id: conn_script
report:
s3_params:
bucket: prueba-id
......
deploy-k8/requirements.txt
View file @ b4835849
openpyxl==3.1.2
XlsxWriter==3.1.2
pymysql==1.1.0
oracledb==1.3.2
apache-airflow-providers-google
apache-airflow-providers-amazon
apache-airflow-providers-postgres
apache-airflow-providers-oracle
\ No newline at end of file
pip install openpyxl==3.1.2
pip install XlsxWriter==3.1.2
pip install pymysql==1.1.0
pip install oracledb==1.3.2
pip install apache-airflow-providers-google
pip install apache-airflow-providers-amazon
pip install apache-airflow-providers-postgres
pip install apache-airflow-providers-oracle
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment