PySpark Design Patterns Quick Reference

 1from abc import ABC, abstractmethod
 2
 3class DataSourceFactory(ABC):
 4    @abstractmethod
 5    def create_data_source(self):
 6        pass
 7
 8class CSVFactory(DataSourceFactory):
 9    def create_data_source(self):
10        return CSVDataSource()
11
12class ParquetFactory(DataSourceFactory):
13    def create_data_source(self):
14        return ParquetDataSource()
15
16# Usage
17factory = CSVFactory()
18source = factory.create_data_source()

 1from threading import Lock
 2
 3class SparkConfig:
 4    _instance = None
 5    _lock = Lock()
 6    
 7    def __new__(cls):
 8        with cls._lock:
 9            if cls._instance is None:
10                cls._instance = super().__new__(cls)
11        return cls._instance
12
13# Usage
14config1 = SparkConfig()
15config2 = SparkConfig()
16assert config1 is config2  # Same instance

 1class TransformBuilder:
 2    def __init__(self):
 3        self.filters = []
 4        self.aggregations = {}
 5    
 6    def add_filter(self, condition):
 7        self.filters.append(condition)
 8        return self
 9    
10    def add_aggregation(self, col, func):
11        self.aggregations[col] = func
12        return self
13    
14    def build(self):
15        return DataTransform(self.filters, self.aggregations)
16
17# Usage
18transform = (TransformBuilder()
19    .add_filter("status == 'active'")
20    .add_aggregation("amount", "sum")
21    .build())

 1from abc import ABC, abstractmethod
 2
 3class Observer(ABC):
 4    @abstractmethod
 5    def update(self, event):
 6        pass
 7
 8class DataSubject:
 9    def __init__(self):
10        self.observers = []
11    
12    def register(self, observer):
13        self.observers.append(observer)
14    
15    def notify(self, event):
16        for obs in self.observers:
17            obs.update(event)
18
19# Usage
20subject = DataSubject()
21subject.register(LoggingObserver())
22subject.notify(DataEvent(data))

 1from abc import ABC, abstractmethod
 2
 3class Transform(ABC):
 4    def __init__(self, next_transform=None):
 5        self.next = next_transform
 6    
 7    def set_next(self, transform):
 8        self.next = transform
 9        return transform
10    
11    @abstractmethod
12    def process(self, data):
13        pass
14
15class CleanTransform(Transform):
16    def process(self, data):
17        cleaned = data.dropna()
18        return self.next.process(cleaned) if self.next else cleaned
19
20# Usage
21pipeline = CleanTransform()
22pipeline.set_next(ValidateTransform()).set_next(EnrichTransform())
23result = pipeline.process(df)