def sql_json(self):
"""Runs arbitrary sql and returns and json"""
session = db.session()
limit = 1000
sql = request.form.get('sql')
database_id = request.form.get('database_id')
mydb = session.query(models.Database).filter_by(id=database_id).first()
if not (self.can_access(
'all_datasource_access', 'all_datasource_access') or
self.can_access('database_access', mydb.perm)):
raise utils.CaravelSecurityException(_(
"SQL Lab requires the `all_datasource_access` or "
"specific DB permission"))
error_msg = ""
if not mydb:
error_msg = "The database selected doesn't seem to exist"
else:
eng = mydb.get_sqla_engine()
if limit:
sql = sql.strip().strip(';')
qry = (
select('*')
.select_from(TextAsFrom(text(sql), ['*'])
.alias('inner_qry'))
.limit(limit)
)
sql = '{}'.format(qry.compile(
eng, compile_kwargs={"literal_binds": True}))
try:
df = pd.read_sql_query(sql=sql, con=eng)
df = df.fillna(0) # TODO make sure NULL
except Exception as e:
logging.exception(e)
error_msg = utils.error_msg_from_exception(e)
session.commit()
if error_msg:
return Response(
json.dumps({
'error': error_msg,
}),
status=500,
mimetype="application/json")
else:
data = {
'columns': [c for c in df.columns],
'data': df.to_dict(orient='records'),
}
return json.dumps(
data, default=utils.json_int_dttm_ser, allow_nan=False)
def refresh_datasources(self):
"""endpoint that refreshes druid datasources metadata"""
session = db.session()
for cluster in session.query(models.DruidCluster).all():
cluster_name = cluster.cluster_name
try:
cluster.refresh_datasources()
except Exception as e:
flash(
"Error while processing cluster '{}'\n{}".format(
cluster_name, utils.error_msg_from_exception(e)),
"danger")
logging.exception(e)
return redirect('/druidclustermodelview/list/')
cluster.metadata_last_refreshed = datetime.now()
flash(
"Refreshed metadata from cluster "
"[" + cluster.cluster_name + "]",
'info')
session.commit()
return redirect("/druiddatasourcemodelview/list/")
def get_sql_results(self, query_id, return_results=True, store_results=False):
"""Executes the sql query returns the results."""
if not self.request.called_directly:
engine = sqlalchemy.create_engine(
app.config.get('SQLALCHEMY_DATABASE_URI'), poolclass=NullPool)
session_class = sessionmaker()
session_class.configure(bind=engine)
session = session_class()
else:
session = db.session()
session.commit() # HACK
query = session.query(models.Query).filter_by(id=query_id).one()
database = query.database
executed_sql = query.sql.strip().strip(';')
db_engine_spec = database.db_engine_spec
def handle_error(msg):
"""Local method handling error while processing the SQL"""
query.error_message = msg
query.status = QueryStatus.FAILED
query.tmp_table_name = None
session.commit()
raise Exception(query.error_message)
# Limit enforced only for retrieving the data, not for the CTA queries.
is_select = is_query_select(executed_sql)
if not is_select and not database.allow_dml:
handle_error(
"Only `SELECT` statements are allowed against this database")
if query.select_as_cta:
if not is_select:
handle_error(
"Only `SELECT` statements can be used with the CREATE TABLE "
"feature.")
if not query.tmp_table_name:
start_dttm = datetime.fromtimestamp(query.start_time)
query.tmp_table_name = 'tmp_{}_table_{}'.format(
query.user_id, start_dttm.strftime('%Y_%m_%d_%H_%M_%S'))
executed_sql = create_table_as(executed_sql, query.tmp_table_name,
database.force_ctas_schema)
query.select_as_cta_used = True
elif (query.limit and is_select
and db_engine_spec.limit_method == LimitMethod.WRAP_SQL):
executed_sql = database.wrap_sql_limit(executed_sql, query.limit)
query.limit_used = True
engine = database.get_sqla_engine(schema=query.schema)
try:
template_processor = get_template_processor(database=database,
query=query)
executed_sql = template_processor.process_template(executed_sql)
except Exception as e:
logging.exception(e)
msg = "Template rendering failed: " + utils.error_msg_from_exception(e)
handle_error(msg)
try:
query.executed_sql = executed_sql
logging.info("Running query: \n{}".format(executed_sql))
result_proxy = engine.execute(query.executed_sql, schema=query.schema)
except Exception as e:
logging.exception(e)
handle_error(utils.error_msg_from_exception(e))
cursor = result_proxy.cursor
query.status = QueryStatus.RUNNING
session.flush()
db_engine_spec.handle_cursor(cursor, query, session)
cdf = None
if result_proxy.cursor:
column_names = [col[0] for col in result_proxy.cursor.description]
if db_engine_spec.limit_method == LimitMethod.FETCH_MANY:
data = result_proxy.fetchmany(query.limit)
else:
data = result_proxy.fetchall()
cdf = dataframe.CaravelDataFrame(
pd.DataFrame(data, columns=column_names))
query.rows = result_proxy.rowcount
query.progress = 100
query.status = QueryStatus.SUCCESS
if query.rows == -1 and cdf:
# Presto doesn't provide result_proxy.row_count
query.rows = cdf.size
if query.select_as_cta:
query.select_sql = '{}'.format(
database.select_star(query.tmp_table_name, limit=query.limit))
query.end_time = utils.now_as_float()
session.flush()
payload = {
'query_id': query.id,
'status': query.status,
'data': [],
}
payload['data'] = cdf.data if cdf else []
payload['columns'] = cdf.columns_dict if cdf else []
payload['query'] = query.to_dict()
payload = json.dumps(payload, default=utils.json_iso_dttm_ser)
if store_results and results_backend:
key = '{}'.format(uuid.uuid4())
logging.info("Storing results in results backend, key: {}".format(key))
results_backend.set(key, zlib.compress(payload))
query.results_key = key
session.flush()
session.commit()
if return_results:
return payload
def get_sql_results(query_id, return_results=True):
"""Executes the sql query returns the results."""
session = db.session()
session.commit() # HACK
query = session.query(models.Query).filter_by(id=query_id).one()
database = query.database
executed_sql = query.sql.strip().strip(';')
def handle_error(msg):
"""Local method handling error while processing the SQL"""
query.error_message = msg
query.status = QueryStatus.FAILED
query.tmp_table_name = None
session.commit()
raise Exception(query.error_message)
# Limit enforced only for retrieving the data, not for the CTA queries.
is_select = is_query_select(executed_sql);
if not is_select and not database.allow_dml:
handle_error(
"Only `SELECT` statements are allowed against this database")
if query.select_as_cta:
if not is_select:
handle_error(
"Only `SELECT` statements can be used with the CREATE TABLE "
"feature.")
if not query.tmp_table_name:
start_dttm = datetime.fromtimestamp(query.start_time)
query.tmp_table_name = 'tmp_{}_table_{}'.format(
query.user_id,
start_dttm.strftime('%Y_%m_%d_%H_%M_%S'))
executed_sql = create_table_as(
executed_sql, query.tmp_table_name, database.force_ctas_schema)
query.select_as_cta_used = True
elif query.limit and is_select:
executed_sql = database.wrap_sql_limit(executed_sql, query.limit)
query.limit_used = True
engine = database.get_sqla_engine(schema=query.schema)
try:
query.executed_sql = executed_sql
logging.info("Running query: \n{}".format(executed_sql))
result_proxy = engine.execute(query.executed_sql, schema=query.schema)
except Exception as e:
logging.exception(e)
handle_error(utils.error_msg_from_exception(e))
cursor = result_proxy.cursor
query.status = QueryStatus.RUNNING
session.flush()
if database.backend == 'presto':
polled = cursor.poll()
# poll returns dict -- JSON status information or ``None``
# if the query is done
# https://github.com/dropbox/PyHive/blob/
# b34bdbf51378b3979eaf5eca9e956f06ddc36ca0/pyhive/presto.py#L178
while polled:
# Update the object and wait for the kill signal.
stats = polled.get('stats', {})
if stats:
completed_splits = float(stats.get('completedSplits'))
total_splits = float(stats.get('totalSplits'))
if total_splits and completed_splits:
progress = 100 * (completed_splits / total_splits)
if progress > query.progress:
query.progress = progress
session.commit()
time.sleep(1)
polled = cursor.poll()
cdf = None
if result_proxy.cursor:
column_names = [col[0] for col in result_proxy.cursor.description]
data = result_proxy.fetchall()
cdf = dataframe.CaravelDataFrame(
pd.DataFrame(data, columns=column_names))
# TODO consider generating tuples instead of dicts to send
# less data through the wire. The command bellow does that,
# but we'd need to align on the client side.
# data = df.values.tolist()
query.rows = result_proxy.rowcount
query.progress = 100
query.status = QueryStatus.SUCCESS
if query.rows == -1 and cdf:
# Presto doesn't provide result_proxy.row_count
query.rows = cdf.size
if query.select_as_cta:
query.select_sql = '{}'.format(database.select_star(
query.tmp_table_name, limit=query.limit))
query.end_time = utils.now_as_float()
session.commit()
if return_results:
payload = {
'query_id': query.id,
'status': query.status,
'data': [],
}
if query.status == models.QueryStatus.SUCCESS:
payload['data'] = cdf.data if cdf else []
payload['columns'] = cdf.columns_dict if cdf else []
else:
payload['error'] = query.error_message
return payload
'''
def get_sql_results(query_id, return_results=True):
"""Executes the sql query returns the results."""
db.session.commit() # HACK
q = db.session.query(models.Query).all()
query = db.session.query(models.Query).filter_by(id=query_id).one()
database = query.database
executed_sql = query.sql.strip().strip(';')
# Limit enforced only for retrieving the data, not for the CTA queries.
if is_query_select(executed_sql):
if query.select_as_cta:
if not query.tmp_table_name:
start_dttm = datetime.fromtimestamp(query.start_time)
query.tmp_table_name = 'tmp_{}_table_{}'.format(
query.user_id, start_dttm.strftime('%Y_%m_%d_%H_%M_%S'))
executed_sql = create_table_as(executed_sql, query.tmp_table_name)
query.select_as_cta_used = True
elif query.limit:
executed_sql = database.wrap_sql_limit(executed_sql, query.limit)
query.limit_used = True
engine = database.get_sqla_engine(schema=query.schema)
try:
query.executed_sql = executed_sql
logging.info("Running query: \n{}".format(executed_sql))
result_proxy = engine.execute(query.executed_sql,
schema=query.schema)
except Exception as e:
logging.exception(e)
query.error_message = utils.error_msg_from_exception(e)
query.status = QueryStatus.FAILED
query.tmp_table_name = None
db.session.commit()
raise Exception(query.error_message)
cursor = result_proxy.cursor
query.status = QueryStatus.RUNNING
db.session.flush()
if hasattr(cursor, "poll"):
polled = cursor.poll()
# poll returns dict -- JSON status information or ``None``
# if the query is done
# https://github.com/dropbox/PyHive/blob/
# b34bdbf51378b3979eaf5eca9e956f06ddc36ca0/pyhive/presto.py#L178
while polled:
# Update the object and wait for the kill signal.
stats = polled.get('stats', {})
if stats:
completed_splits = float(stats.get('completedSplits'))
total_splits = float(stats.get('totalSplits'))
if total_splits and completed_splits:
progress = 100 * (completed_splits / total_splits)
if progress > query.progress:
query.progress = progress
db.session.commit()
time.sleep(1)
polled = cursor.poll()
columns = None
data = None
if result_proxy.cursor:
cols = [col[0] for col in result_proxy.cursor.description]
data = result_proxy.fetchall()
df = pd.DataFrame(data, columns=cols)
columns = [c for c in df.columns]
data = df.to_dict(orient='records')
query.rows = result_proxy.rowcount
query.progress = 100
query.status = QueryStatus.SUCCESS
if query.rows == -1 and data:
# Presto doesn't provide result_proxy.row_count
query.rows = len(data)
# CTAs queries result in 1 cell having the # of the added rows.
if query.select_as_cta:
query.select_sql = '{}'.format(
database.select_star(query.tmp_table_name, limit=query.limit))
query.end_time = utils.now_as_float()
db.session.commit()
payload = {
'query_id': query.id,
'status': query.status,
}
if query.status == models.QueryStatus.SUCCESS:
payload['data'] = data
payload['columns'] = columns
else:
payload['error'] = query.error_message
if return_results:
return payload
def get_sql_results(query_id, return_results=True):
"""Executes the sql query returns the results."""
session = db.session()
session.commit() # HACK
query = session.query(models.Query).filter_by(id=query_id).one()
database = query.database
executed_sql = query.sql.strip().strip(';')
def handle_error(msg):
"""Local method handling error while processing the SQL"""
query.error_message = msg
query.status = QueryStatus.FAILED
query.tmp_table_name = None
session.commit()
raise Exception(query.error_message)
# Limit enforced only for retrieving the data, not for the CTA queries.
is_select = is_query_select(executed_sql)
if not is_select and not database.allow_dml:
handle_error(
"Only `SELECT` statements are allowed against this database")
if query.select_as_cta:
if not is_select:
handle_error(
"Only `SELECT` statements can be used with the CREATE TABLE "
"feature.")
if not query.tmp_table_name:
start_dttm = datetime.fromtimestamp(query.start_time)
query.tmp_table_name = 'tmp_{}_table_{}'.format(
query.user_id, start_dttm.strftime('%Y_%m_%d_%H_%M_%S'))
executed_sql = create_table_as(executed_sql, query.tmp_table_name,
database.force_ctas_schema)
query.select_as_cta_used = True
elif query.limit and is_select:
executed_sql = database.wrap_sql_limit(executed_sql, query.limit)
query.limit_used = True
engine = database.get_sqla_engine(schema=query.schema)
try:
query.executed_sql = executed_sql
logging.info("Running query: \n{}".format(executed_sql))
result_proxy = engine.execute(query.executed_sql, schema=query.schema)
except Exception as e:
logging.exception(e)
handle_error(utils.error_msg_from_exception(e))
cursor = result_proxy.cursor
query.status = QueryStatus.RUNNING
session.flush()
if database.backend == 'presto':
polled = cursor.poll()
# poll returns dict -- JSON status information or ``None``
# if the query is done
# https://github.com/dropbox/PyHive/blob/
# b34bdbf51378b3979eaf5eca9e956f06ddc36ca0/pyhive/presto.py#L178
while polled:
# Update the object and wait for the kill signal.
stats = polled.get('stats', {})
if stats:
completed_splits = float(stats.get('completedSplits'))
total_splits = float(stats.get('totalSplits'))
if total_splits and completed_splits:
progress = 100 * (completed_splits / total_splits)
if progress > query.progress:
query.progress = progress
session.commit()
time.sleep(1)
polled = cursor.poll()
columns = None
data = None
if result_proxy.cursor:
columns = [col[0] for col in result_proxy.cursor.description]
data = result_proxy.fetchall()
df = pd.DataFrame(data, columns=columns)
df = df.where((pd.notnull(df)), None)
# TODO consider generating tuples instead of dicts to send
# less data through the wire. The command bellow does that,
# but we'd need to align on the client side.
# data = df.values.tolist()
data = df.to_dict(orient='records')
query.rows = result_proxy.rowcount
query.progress = 100
query.status = QueryStatus.SUCCESS
if query.rows == -1 and data:
# Presto doesn't provide result_proxy.row_count
query.rows = len(data)
# CTAs queries result in 1 cell having the # of the added rows.
if query.select_as_cta:
query.select_sql = '{}'.format(
database.select_star(query.tmp_table_name, limit=query.limit))
query.end_time = utils.now_as_float()
session.commit()
payload = {
'query_id': query.id,
'status': query.status,
'data': [],
}
if query.status == models.QueryStatus.SUCCESS:
payload['data'] = data
payload['columns'] = columns
else:
payload['error'] = query.error_message
if return_results:
return payload
'''
def get_sql_results(self, query_id, return_results=True, store_results=False):
"""Executes the sql query returns the results."""
if not self.request.called_directly:
engine = sqlalchemy.create_engine(
app.config.get('SQLALCHEMY_DATABASE_URI'), poolclass=NullPool)
session_class = sessionmaker()
session_class.configure(bind=engine)
session = session_class()
else:
session = db.session()
session.commit() # HACK
query = session.query(models.Query).filter_by(id=query_id).one()
database = query.database
executed_sql = query.sql.strip().strip(';')
db_engine_spec = database.db_engine_spec
def handle_error(msg):
"""Local method handling error while processing the SQL"""
query.error_message = msg
query.status = QueryStatus.FAILED
query.tmp_table_name = None
session.commit()
raise Exception(query.error_message)
# Limit enforced only for retrieving the data, not for the CTA queries.
is_select = is_query_select(executed_sql);
if not is_select and not database.allow_dml:
handle_error(
"Only `SELECT` statements are allowed against this database")
if query.select_as_cta:
if not is_select:
handle_error(
"Only `SELECT` statements can be used with the CREATE TABLE "
"feature.")
if not query.tmp_table_name:
start_dttm = datetime.fromtimestamp(query.start_time)
query.tmp_table_name = 'tmp_{}_table_{}'.format(
query.user_id,
start_dttm.strftime('%Y_%m_%d_%H_%M_%S'))
executed_sql = create_table_as(
executed_sql, query.tmp_table_name, database.force_ctas_schema)
query.select_as_cta_used = True
elif (
query.limit and is_select and
db_engine_spec.limit_method == LimitMethod.WRAP_SQL):
executed_sql = database.wrap_sql_limit(executed_sql, query.limit)
query.limit_used = True
engine = database.get_sqla_engine(schema=query.schema)
try:
template_processor = get_template_processor(
database=database, query=query)
executed_sql = template_processor.process_template(executed_sql)
except Exception as e:
logging.exception(e)
msg = "Template rendering failed: " + utils.error_msg_from_exception(e)
handle_error(msg)
try:
query.executed_sql = executed_sql
logging.info("Running query: \n{}".format(executed_sql))
result_proxy = engine.execute(query.executed_sql, schema=query.schema)
except Exception as e:
logging.exception(e)
handle_error(utils.error_msg_from_exception(e))
cursor = result_proxy.cursor
query.status = QueryStatus.RUNNING
session.flush()
db_engine_spec.handle_cursor(cursor, query, session)
cdf = None
if result_proxy.cursor:
column_names = [col[0] for col in result_proxy.cursor.description]
if db_engine_spec.limit_method == LimitMethod.FETCH_MANY:
data = result_proxy.fetchmany(query.limit)
else:
data = result_proxy.fetchall()
cdf = dataframe.CaravelDataFrame(
pd.DataFrame(data, columns=column_names))
query.rows = result_proxy.rowcount
query.progress = 100
query.status = QueryStatus.SUCCESS
if query.rows == -1 and cdf:
# Presto doesn't provide result_proxy.row_count
query.rows = cdf.size
if query.select_as_cta:
query.select_sql = '{}'.format(database.select_star(
query.tmp_table_name, limit=query.limit))
query.end_time = utils.now_as_float()
session.flush()
payload = {
'query_id': query.id,
'status': query.status,
'data': [],
}
payload['data'] = cdf.data if cdf else []
payload['columns'] = cdf.columns_dict if cdf else []
payload['query'] = query.to_dict()
payload = json.dumps(payload, default=utils.json_iso_dttm_ser)
if store_results and results_backend:
key = '{}'.format(uuid.uuid4())
logging.info("Storing results in results backend, key: {}".format(key))
results_backend.set(key, zlib.compress(payload))
query.results_key = key
session.flush()
session.commit()
if return_results:
return payload
def get_sql_results(query_id, return_results=True):
"""Executes the sql query returns the results."""
db.session.commit() # HACK
query = db.session.query(models.Query).filter_by(id=query_id).one()
database = query.database
executed_sql = query.sql.strip().strip(';')
# Limit enforced only for retrieving the data, not for the CTA queries.
if is_query_select(executed_sql):
if query.select_as_cta:
if not query.tmp_table_name:
start_dttm = datetime.fromtimestamp(query.start_time)
query.tmp_table_name = 'tmp_{}_table_{}'.format(
query.user_id,
start_dttm.strftime('%Y_%m_%d_%H_%M_%S'))
executed_sql = create_table_as(
executed_sql, query.tmp_table_name, database.force_ctas_schema)
query.select_as_cta_used = True
elif query.limit:
executed_sql = database.wrap_sql_limit(executed_sql, query.limit)
query.limit_used = True
engine = database.get_sqla_engine(schema=query.schema)
try:
query.executed_sql = executed_sql
logging.info("Running query: \n{}".format(executed_sql))
result_proxy = engine.execute(query.executed_sql, schema=query.schema)
except Exception as e:
logging.exception(e)
query.error_message = utils.error_msg_from_exception(e)
query.status = QueryStatus.FAILED
query.tmp_table_name = None
db.session.commit()
raise Exception(query.error_message)
cursor = result_proxy.cursor
query.status = QueryStatus.RUNNING
db.session.flush()
if database.backend == 'presto':
polled = cursor.poll()
# poll returns dict -- JSON status information or ``None``
# if the query is done
# https://github.com/dropbox/PyHive/blob/
# b34bdbf51378b3979eaf5eca9e956f06ddc36ca0/pyhive/presto.py#L178
while polled:
# Update the object and wait for the kill signal.
stats = polled.get('stats', {})
if stats:
completed_splits = float(stats.get('completedSplits'))
total_splits = float(stats.get('totalSplits'))
if total_splits and completed_splits:
progress = 100 * (completed_splits / total_splits)
if progress > query.progress:
query.progress = progress
db.session.commit()
time.sleep(1)
polled = cursor.poll()
columns = None
data = None
if result_proxy.cursor:
columns = [col[0] for col in result_proxy.cursor.description]
data = result_proxy.fetchall()
df = pd.DataFrame(data, columns=columns)
df = df.where((pd.notnull(df)), None)
# TODO consider generating tuples instead of dicts to send
# less data through the wire. The command bellow does that,
# but we'd need to align on the client side.
# data = df.values.tolist()
data = df.to_dict(orient='records')
query.rows = result_proxy.rowcount
query.progress = 100
query.status = QueryStatus.SUCCESS
if query.rows == -1 and data:
# Presto doesn't provide result_proxy.row_count
query.rows = len(data)
# CTAs queries result in 1 cell having the # of the added rows.
if query.select_as_cta:
query.select_sql = '{}'.format(database.select_star(
query.tmp_table_name, limit=query.limit))
query.end_time = utils.now_as_float()
db.session.commit()
payload = {
'query_id': query.id,
'status': query.status,
}
if query.status == models.QueryStatus.SUCCESS:
payload['data'] = data
payload['columns'] = columns
else:
payload['error'] = query.error_message
if return_results:
return payload