brief description
PyFlink is Apache Flink 's Python API that you can use to build scalable batch and stream processing tasks, such as real-time data processing pipelines, large-scale exploratory data analysis, machine learning ( ML ) pipelines, and ETL processing.
If you are already familiar with libraries such as Python and Pandas , then PyFlink makes it easier for you to take advantage of the full capabilities of the Flink ecosystem.
Depending on the level of abstraction you need, there are two different APIs that can be used in PyFlink :
| PyFlink Table API | Allows you to write powerful relational queries in a manner similar to SQL or to tabular data in Python . |
| PyFlink DataStream API | Allows you to have fine-grained control over state and time , the core components of Flink , in order to build more complex stream processing applications. |
Table API
Apache Flink provides the Table API relational API to unify processing streams and batches, that is, queries are executed with the same semantics on unbounded real-time streams or bounded batch datasets, and produce the same results.
Flink's Table API is easy to write and generally simplifies the coding of data analysis, data pipelines, and ETL applications.
All Table API and SQL programs, regardless of batch mode or streaming mode, follow the same structure.
We will start from scratch and introduce how to create a Flink Python project and run Python Table API jobs.
This job reads a csv file, calculates word frequencies, and writes the results to a results file.
code example
WorldCount.py
import argparse
import logging
import sys
from pyflink.common import Row
from pyflink.table import (Environmentsettings, TableEnvironment, TableDescriptor, Schema,DataTypes, FormatDescriptor)
from pyflink.table.expressions import lit, col
from pyflink.table.udf import udtf
word_count_data = ["To be, or not to be,--that is the question:--",
"Whether 'tis nobler in the mind to suffer",
"The slings and arrows of outrageous fortune",
"Or to take arms against a sea of troubles,",
"And by opposing end them?--To die,--to sleep,--",
"No more; and by a sleep to say we end",
"The heartache, and the thousand natural shocks",
"That flesh is heir to,--'tis a consummation",
"Devoutly to be wish'd. To die,--to sleep;--",
"To sleep! perchance to dream:--ay, there's the rub;",
"For in that sleep of death what dreams may come,",
"When we have shuffled off this mortal coil,",
"Must give us pause: there's the respect",
"That makes calamity of so long life;",
"For who would bear the whips and scorns of time,",
"The oppressor's wrong, the proud man's contumely,",
"The pangs of despis'd love, the law's delay,",
"The insolence of office, and the spurns",
"That patient merit of the unworthy takes,",
"When he himself might his quietus make",
"With a bare bodkin? who would these fardels bear,",
"To grunt and sweat under a weary life,",
"But that the dread of something after death,--",
"The undiscover'd country, from whose bourn",
"No traveller returns,--puzzles the will,",
"And makes us rather bear those ills we have",
"Than fly to others that we know not of?",
"Thus conscience does make cowards of us all;",
"And thus the native hue of resolution",
"Is sicklied o'er with the pale cast of thought;",
"And enterprises of great pith and moment,",
"With this regard, their currents turn awry,",
"And lose the name of action.--Soft you now!",
"The fair Ophelia!--Nymph, in thy orisons",
"Be all my sins remember'd."]
def word_count(input_path, output_path):
t_env = TableEnvironment.create(Environmentsettings.in_streaming_mode())
# write all the data to one file
t_env.get_config().get_configuration().set_string("parallelism.default", "1")
# define the source
if input_path is not None:
t_env.create_temporary_table(
'source',
TableDescriptor.for_connector('filesystem')
.schema(Schema.new_builder()
.column('word', DataTypes.STRING())
.build())
.option('path', input_path)
.format('csv')
.build())
tab = t_env.from_path('source')
else:
print("Executing word_count example with default input data set.")
print("Use --input to specify file input.")
tab = t_env.from_elements(map(lambda i: (i,), word_count_data),
DataTypes.ROW([DataTypes.FIELD('line', DataTypes.STRING())]))
# define the sink
if output_path is not None:
t_env.create_temporary_table(
'sink',
TableDescriptor.for_connector('filesystem')
.schema(Schema.new_builder()
.column('word', DataTypes.STRING())
.column('count', DataTypes.BIGINT())
.build())
.option('path', output_path)
.format(FormatDescriptor.for_format('canal-json')
.build())
.build())
else:
print("Printing result to stdout. Use --output to specify output path.")
t_env.create_temporary_table(
'sink',
TableDescriptor.for_connector('print')
.schema(Schema.new_builder()
.column('word', DataTypes.STRING())
.column('count', DataTypes.BIGINT())
.build())
.build())
@udtf(result_types=[DataTypes.STRING()])
def split(line: Row):
for s in line[0].split():
yield Row(s)
# compute word count
tab.flat_map(split).alias('word') \
.group_by(col('word')) \
.select(col('word'), lit(1).count) \
.execute_insert('sink') \
.wait()
# remove .wait if submitting to a remote cluster
if __name__ == '__main__':
logging.basicConfig(stream=sys.stdout, level=logging.INFO, format="%(message)s")
parser = argparse.ArgumentParser()
parser.add_argument(
'--input',
dest='input',
required=False,
help='Input file to process.')
parser.add_argument(
'--output',
dest='output',
required=False,
help='Output file to write results to.')
argv = sys.argv[1:]
known_args, _ = parser.parse_known_args(argv)
word_count(known_args.input, known_args.output)