Source code for airflow.macros.hive
#
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
import datetime
[docs]def max_partition(
table, schema="default", field=None, filter_map=None, metastore_conn_id='metastore_default'
):
"""
Gets the max partition for a table.
:param schema: The hive schema the table lives in
:type schema: str
:param table: The hive table you are interested in, supports the dot
notation as in "my_database.my_table", if a dot is found,
the schema param is disregarded
:type table: str
:param metastore_conn_id: The hive connection you are interested in.
If your default is set you don't need to use this parameter.
:type metastore_conn_id: str
:param filter_map: partition_key:partition_value map used for partition filtering,
e.g. {'key1': 'value1', 'key2': 'value2'}.
Only partitions matching all partition_key:partition_value
pairs will be considered as candidates of max partition.
:type filter_map: dict
:param field: the field to get the max value from. If there's only
one partition field, this will be inferred
:type field: str
>>> max_partition('airflow.static_babynames_partitioned')
'2015-01-01'
"""
from airflow.providers.apache.hive.hooks.hive import HiveMetastoreHook
if '.' in table:
schema, table = table.split('.')
hive_hook = HiveMetastoreHook(metastore_conn_id=metastore_conn_id)
return hive_hook.max_partition(schema=schema, table_name=table, field=field, filter_map=filter_map)
def _closest_date(target_dt, date_list, before_target=None):
"""
This function finds the date in a list closest to the target date.
An optional parameter can be given to get the closest before or after.
:param target_dt: The target date
:type target_dt: datetime.date
:param date_list: The list of dates to search
:type date_list: list[datetime.date]
:param before_target: closest before or after the target
:type before_target: bool or None
:returns: The closest date
:rtype: datetime.date or None
"""
time_before = lambda d: target_dt - d if d <= target_dt else datetime.timedelta.max
time_after = lambda d: d - target_dt if d >= target_dt else datetime.timedelta.max
any_time = lambda d: target_dt - d if d < target_dt else d - target_dt
if before_target is None:
return min(date_list, key=any_time).date()
if before_target:
return min(date_list, key=time_before).date()
else:
return min(date_list, key=time_after).date()
[docs]def closest_ds_partition(table, ds, before=True, schema="default", metastore_conn_id='metastore_default'):
"""
This function finds the date in a list closest to the target date.
An optional parameter can be given to get the closest before or after.
:param table: A hive table name
:type table: str
:param ds: A datestamp ``%Y-%m-%d`` e.g. ``yyyy-mm-dd``
:type ds: list[datetime.date]
:param before: closest before (True), after (False) or either side of ds
:type before: bool or None
:param schema: table schema
:type schema: str
:param metastore_conn_id: which metastore connection to use
:type metastore_conn_id: str
:returns: The closest date
:rtype: str or None
>>> tbl = 'airflow.static_babynames_partitioned'
>>> closest_ds_partition(tbl, '2015-01-02')
'2015-01-01'
"""
from airflow.providers.apache.hive.hooks.hive import HiveMetastoreHook
if '.' in table:
schema, table = table.split('.')
hive_hook = HiveMetastoreHook(metastore_conn_id=metastore_conn_id)
partitions = hive_hook.get_partitions(schema=schema, table_name=table)
if not partitions:
return None
part_vals = [list(p.values())[0] for p in partitions]
if ds in part_vals:
return ds
else:
parts = [datetime.datetime.strptime(pv, '%Y-%m-%d') for pv in part_vals]
target_dt = datetime.datetime.strptime(ds, '%Y-%m-%d')
closest_ds = _closest_date(target_dt, parts, before_target=before)
return closest_ds.isoformat()