datapipelab 0.2.2__tar.gz → 0.2.4__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- {datapipelab-0.2.2 → datapipelab-0.2.4}/PKG-INFO +1 -1
- datapipelab-0.2.4/datapipelab/app/node/sink/spark_node.py +27 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/pipeline_handler.py +4 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab.egg-info/PKG-INFO +1 -1
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab.egg-info/SOURCES.txt +1 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/setup.py +1 -1
- {datapipelab-0.2.2 → datapipelab-0.2.4}/MANIFEST.in +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/README.md +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/__init__.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/__init__.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/connector_node/__init__.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/__init__.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/custom_node.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/processor/__init__.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/processor/bigquery_api_node.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/processor/bigquery_spark_node.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/processor/custom_node.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/processor/shell_node.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/processor/spark_node.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/sink/__init__.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/sink/csv_node.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/sink/delta_node.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/sink/hive_node.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/sink/pandas_csv_node.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/sink/teams_notification_node.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/source/__init__.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/source/delta_node.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/source/hive_node.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/source/spark_node.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/tnode.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/engine.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/logger.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/pipeline.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/pipeline_config.py +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab.egg-info/dependency_links.txt +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab.egg-info/requires.txt +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab.egg-info/top_level.txt +0 -0
- {datapipelab-0.2.2 → datapipelab-0.2.4}/setup.cfg +0 -0
@@ -0,0 +1,27 @@
|
|
1
|
+
from datapipelab.app.node.tnode import TNode
|
2
|
+
from datapipelab.logger import logger
|
3
|
+
|
4
|
+
class SparkSinkNode(TNode):
|
5
|
+
def __init__(self, spark, tnode_config, df):
|
6
|
+
from pyspark.sql import DataFrame
|
7
|
+
super().__init__(spark=spark)
|
8
|
+
self.mode = tnode_config.get('mode', None)
|
9
|
+
self.stream = tnode_config.get('stream', None)
|
10
|
+
self.database_name = tnode_config['options']['database']
|
11
|
+
self.table_name = tnode_config['options']['table']
|
12
|
+
self.partition_by = tnode_config['options'].get('partition_by', None)
|
13
|
+
self.partition_count = tnode_config['options'].get('partition_count', None)
|
14
|
+
self.overwrite = tnode_config['options']['overwrite']
|
15
|
+
self.df = df
|
16
|
+
|
17
|
+
def __write_dynamic_partition(self):
|
18
|
+
if self.partition_count:
|
19
|
+
if self.partition_by:
|
20
|
+
self.df = self.df.repartition(int(self.partition_count))
|
21
|
+
else:
|
22
|
+
self.df = self.df.repartition(int(self.partition_count), self.partition_by)
|
23
|
+
logger.info("Start writing to Hive")
|
24
|
+
self.df.write.insertInto(f'{self.database_name}.{self.table_name}', overwrite=self.overwrite)
|
25
|
+
|
26
|
+
def _process(self):
|
27
|
+
self.__write_dynamic_partition()
|
@@ -88,6 +88,10 @@ class PipelineHandler:
|
|
88
88
|
if tnode_format == 'hive':
|
89
89
|
from datapipelab.app.node.sink import hive_node
|
90
90
|
processor_df = hive_node.HiveSinkNode(self.spark, tnode_config, t_df[tnode_name_df]).run()
|
91
|
+
if tnode_type == 'spark':
|
92
|
+
if tnode_format == 'spark':
|
93
|
+
from datapipelab.app.node.sink import spark_node
|
94
|
+
processor_df = spark_node.SparkSinkNode(self.spark, tnode_config, t_df[tnode_name_df]).run()
|
91
95
|
|
92
96
|
|
93
97
|
|
@@ -28,6 +28,7 @@ datapipelab/app/node/sink/csv_node.py
|
|
28
28
|
datapipelab/app/node/sink/delta_node.py
|
29
29
|
datapipelab/app/node/sink/hive_node.py
|
30
30
|
datapipelab/app/node/sink/pandas_csv_node.py
|
31
|
+
datapipelab/app/node/sink/spark_node.py
|
31
32
|
datapipelab/app/node/sink/teams_notification_node.py
|
32
33
|
datapipelab/app/node/source/__init__.py
|
33
34
|
datapipelab/app/node/source/delta_node.py
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
{datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/processor/bigquery_spark_node.py
RENAMED
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
{datapipelab-0.2.2 → datapipelab-0.2.4}/datapipelab/app/node/sink/teams_notification_node.py
RENAMED
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|